kaggle_smoking_classification.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "toc_visible": true,
      "authorship_tag": "ABX9TyNdtCj1AGLFxOHYGYYCZxaa",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/mark-narusov/kaggle_leopard_challenge/blob/main/kaggle_smoking_classification.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Введение"
      ],
      "metadata": {
        "id": "V8ZfEszyDMrj"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Задача — предсказать, курит человек или нет, по его клиническим показаниям. Соответственно, перед нами задача обучения с учителем, а именно — бинарной классификации."
      ],
      "metadata": {
        "id": "glYJtwYJtrrP"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "\n",
        "\n",
        "Целевая метрика [этого соревнования](https://www.kaggle.com/competitions/leopard-challenge-classification/overview) — *F1* — зависит от порога классификации. По советам Глеба в курсе [Data Science с Глебом Михайловым](https://stepik.org/course/113596/syllabus) я сначала подбирал модели по метрике *ROC-AUC*, отражающаю качество модели по всем возможным порогам, и только потом побирал оптимальный порог классификации. И модель, и порог я находил на кросс-валидации с пятью фолдами.\n",
        "\n",
        "Только с *CatBoost*-ом и интерпретируемыми фичами я дошёл до скора в **0.455**, третье место на лидерборде. Чтобы вырваться в топ-один я усилил модель и сблендил с ней вероятностные предсказания случайного леса, натренированном на $100$+ относительно сильных интерактивных фичах (*interaction features*). По ходу обнаружил, что такой случайный лес выдаёт результат лучше на тестовой выборке, поэтому в конечной модели отдавал бОльший вес предсказаниям случайного леса, однако оставлял возможность классифицировать в пользу *CatBoost* если последний вдвое \"более уверен\".  \n",
        "\n",
        "В итоге мне удалось добиться скора **0.479** и закрепиться на первом месте."
      ],
      "metadata": {
        "id": "zsHKtCalX1ef"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Содержание"
      ],
      "metadata": {
        "id": "Qil99NgttmuY"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        ">[Введение](#folderId=1BcJ_ObSwBJvLD2oWyW-kDdL4rRzrdvHh&updateTitle=true&scrollTo=V8ZfEszyDMrj)\n",
        "\n",
        ">[Импортирование библиотек](#folderId=1BcJ_ObSwBJvLD2oWyW-kDdL4rRzrdvHh&updateTitle=true&scrollTo=fEDJ02uvCIlZ)\n",
        "\n",
        ">[Знакомимся с данными](#folderId=1BcJ_ObSwBJvLD2oWyW-kDdL4rRzrdvHh&updateTitle=true&scrollTo=1ZIRmCA3cQ8Q)\n",
        "\n",
        ">[Анализ и создание признаков](#folderId=1BcJ_ObSwBJvLD2oWyW-kDdL4rRzrdvHh&updateTitle=true&scrollTo=9RxSqiqaG_Q5)\n",
        "\n",
        ">>[Холестерин, HDL и LDL](#folderId=1BcJ_ObSwBJvLD2oWyW-kDdL4rRzrdvHh&updateTitle=true&scrollTo=H1Yx1rN4E3sR)\n",
        "\n",
        ">>[Mean Arterial Pressure](#folderId=1BcJ_ObSwBJvLD2oWyW-kDdL4rRzrdvHh&updateTitle=true&scrollTo=gyrtJafGkLgG)\n",
        "\n",
        ">>[Dental Health](#folderId=1BcJ_ObSwBJvLD2oWyW-kDdL4rRzrdvHh&updateTitle=true&scrollTo=eBDrpBN2mJ21)\n",
        "\n",
        ">>[Waist-to-Height Ratio](#folderId=1BcJ_ObSwBJvLD2oWyW-kDdL4rRzrdvHh&updateTitle=true&scrollTo=9iNF_8bNndW6)\n",
        "\n",
        ">>[Зрение и слух](#folderId=1BcJ_ObSwBJvLD2oWyW-kDdL4rRzrdvHh&updateTitle=true&scrollTo=tNQi5ZcPq1Lh)\n",
        "\n",
        ">>[Возраст](#folderId=1BcJ_ObSwBJvLD2oWyW-kDdL4rRzrdvHh&updateTitle=true&scrollTo=apaWPKqRsLfv)\n",
        "\n",
        ">>[Вес, индекс массы тела](#folderId=1BcJ_ObSwBJvLD2oWyW-kDdL4rRzrdvHh&updateTitle=true&scrollTo=BXvI5X3PtmlK)\n",
        "\n",
        ">>[Показатели работы печени](#folderId=1BcJ_ObSwBJvLD2oWyW-kDdL4rRzrdvHh&updateTitle=true&scrollTo=6Jz3loc53BdT)\n",
        "\n",
        ">>[Логарифмирование](#folderId=1BcJ_ObSwBJvLD2oWyW-kDdL4rRzrdvHh&updateTitle=true&scrollTo=shbf8yM1z02P)\n",
        "\n",
        ">>[Binning](#folderId=1BcJ_ObSwBJvLD2oWyW-kDdL4rRzrdvHh&updateTitle=true&scrollTo=9nEOxQV1y_sX)\n",
        "\n",
        ">>[Удаляем ненужное](#folderId=1BcJ_ObSwBJvLD2oWyW-kDdL4rRzrdvHh&updateTitle=true&scrollTo=QydIxxaBHF9g)\n",
        "\n",
        ">>[Итог преобразования](#folderId=1BcJ_ObSwBJvLD2oWyW-kDdL4rRzrdvHh&updateTitle=true&scrollTo=2es7urQc5oeR)\n",
        "\n",
        ">[Предобработка для ML](#folderId=1BcJ_ObSwBJvLD2oWyW-kDdL4rRzrdvHh&updateTitle=true&scrollTo=4t-v0VmH5xnv)\n",
        "\n",
        ">>[Предобработка для случайного леса](#folderId=1BcJ_ObSwBJvLD2oWyW-kDdL4rRzrdvHh&updateTitle=true&scrollTo=LSSGbBEKR8jy)\n",
        "\n",
        ">>>[Масштабирование](#folderId=1BcJ_ObSwBJvLD2oWyW-kDdL4rRzrdvHh&updateTitle=true&scrollTo=n1zVopCSGIE-)\n",
        "\n",
        ">>>[Interaction Features](#folderId=1BcJ_ObSwBJvLD2oWyW-kDdL4rRzrdvHh&updateTitle=true&scrollTo=8Ooj3gujGubj)\n",
        "\n",
        ">>>[Выбор признаков](#folderId=1BcJ_ObSwBJvLD2oWyW-kDdL4rRzrdvHh&updateTitle=true&scrollTo=OLX3ThaMIJAx)\n",
        "\n",
        ">>>[OHE-Кодирование категориальных признаков](#folderId=1BcJ_ObSwBJvLD2oWyW-kDdL4rRzrdvHh&updateTitle=true&scrollTo=GgKXTafMJoN9)\n",
        "\n",
        ">[Оптимизация CatBoost-а](#folderId=1BcJ_ObSwBJvLD2oWyW-kDdL4rRzrdvHh&updateTitle=true&scrollTo=atJabIfEF2Gi)\n",
        "\n",
        ">>[Подбор гиперпараметров](#folderId=1BcJ_ObSwBJvLD2oWyW-kDdL4rRzrdvHh&updateTitle=true&scrollTo=0ZyMa3gHPoNb)\n",
        "\n",
        ">>[Подбор оптимального порога](#folderId=1BcJ_ObSwBJvLD2oWyW-kDdL4rRzrdvHh&updateTitle=true&scrollTo=f6G_3Dk1Pg-Y)\n",
        "\n",
        ">[Оптимизация случайного леса](#folderId=1BcJ_ObSwBJvLD2oWyW-kDdL4rRzrdvHh&updateTitle=true&scrollTo=b8EEieJ6M02F)\n",
        "\n",
        ">>[Подбор гиперпараметров](#folderId=1BcJ_ObSwBJvLD2oWyW-kDdL4rRzrdvHh&updateTitle=true&scrollTo=w5ATpGITM02U)\n",
        "\n",
        ">>[Подбор оптимального порога](#folderId=1BcJ_ObSwBJvLD2oWyW-kDdL4rRzrdvHh&updateTitle=true&scrollTo=0MkGulP7M02V)\n",
        "\n",
        ">[Блендинг двух моделей](#folderId=1BcJ_ObSwBJvLD2oWyW-kDdL4rRzrdvHh&updateTitle=true&scrollTo=LAUWkuSOBJY2)\n"
      ],
      "metadata": {
        "id": "Xeh_Tj5stlbS"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Импортирование библиотек"
      ],
      "metadata": {
        "id": "fEDJ02uvCIlZ"
      }
    },
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "uaPzoQ63AsTV",
        "outputId": "b64ae503-8e54-4038-83e3-71fe0c416842"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Requirement already satisfied: phik in /usr/local/lib/python3.10/dist-packages (0.12.3)\n",
            "Requirement already satisfied: ydata_profiling in /usr/local/lib/python3.10/dist-packages (4.3.2)\n",
            "Requirement already satisfied: catboost in /usr/local/lib/python3.10/dist-packages (1.2)\n",
            "Requirement already satisfied: optuna in /usr/local/lib/python3.10/dist-packages (3.2.0)\n",
            "Requirement already satisfied: numpy>=1.18.0 in /usr/local/lib/python3.10/dist-packages (from phik) (1.22.4)\n",
            "Requirement already satisfied: scipy>=1.5.2 in /usr/local/lib/python3.10/dist-packages (from phik) (1.10.1)\n",
            "Requirement already satisfied: pandas>=0.25.1 in /usr/local/lib/python3.10/dist-packages (from phik) (1.5.3)\n",
            "Requirement already satisfied: matplotlib>=2.2.3 in /usr/local/lib/python3.10/dist-packages (from phik) (3.7.1)\n",
            "Requirement already satisfied: joblib>=0.14.1 in /usr/local/lib/python3.10/dist-packages (from phik) (1.3.1)\n",
            "Requirement already satisfied: pydantic<2,>=1.8.1 in /usr/local/lib/python3.10/dist-packages (from ydata_profiling) (1.10.12)\n",
            "Requirement already satisfied: PyYAML<6.1,>=5.0.0 in /usr/local/lib/python3.10/dist-packages (from ydata_profiling) (6.0.1)\n",
            "Requirement already satisfied: jinja2<3.2,>=2.11.1 in /usr/local/lib/python3.10/dist-packages (from ydata_profiling) (3.1.2)\n",
            "Requirement already satisfied: visions[type_image_path]==0.7.5 in /usr/local/lib/python3.10/dist-packages (from ydata_profiling) (0.7.5)\n",
            "Requirement already satisfied: htmlmin==0.1.12 in /usr/local/lib/python3.10/dist-packages (from ydata_profiling) (0.1.12)\n",
            "Requirement already satisfied: requests<3,>=2.24.0 in /usr/local/lib/python3.10/dist-packages (from ydata_profiling) (2.27.1)\n",
            "Requirement already satisfied: tqdm<5,>=4.48.2 in /usr/local/lib/python3.10/dist-packages (from ydata_profiling) (4.65.0)\n",
            "Requirement already satisfied: seaborn<0.13,>=0.10.1 in /usr/local/lib/python3.10/dist-packages (from ydata_profiling) (0.12.2)\n",
            "Requirement already satisfied: multimethod<2,>=1.4 in /usr/local/lib/python3.10/dist-packages (from ydata_profiling) (1.9.1)\n",
            "Requirement already satisfied: statsmodels<1,>=0.13.2 in /usr/local/lib/python3.10/dist-packages (from ydata_profiling) (0.13.5)\n",
            "Requirement already satisfied: typeguard<3,>=2.13.2 in /usr/local/lib/python3.10/dist-packages (from ydata_profiling) (2.13.3)\n",
            "Requirement already satisfied: imagehash==4.3.1 in /usr/local/lib/python3.10/dist-packages (from ydata_profiling) (4.3.1)\n",
            "Requirement already satisfied: wordcloud>=1.9.1 in /usr/local/lib/python3.10/dist-packages (from ydata_profiling) (1.9.2)\n",
            "Requirement already satisfied: dacite>=1.8 in /usr/local/lib/python3.10/dist-packages (from ydata_profiling) (1.8.1)\n",
            "Requirement already satisfied: PyWavelets in /usr/local/lib/python3.10/dist-packages (from imagehash==4.3.1->ydata_profiling) (1.4.1)\n",
            "Requirement already satisfied: pillow in /usr/local/lib/python3.10/dist-packages (from imagehash==4.3.1->ydata_profiling) (9.4.0)\n",
            "Requirement already satisfied: attrs>=19.3.0 in /usr/local/lib/python3.10/dist-packages (from visions[type_image_path]==0.7.5->ydata_profiling) (23.1.0)\n",
            "Requirement already satisfied: networkx>=2.4 in /usr/local/lib/python3.10/dist-packages (from visions[type_image_path]==0.7.5->ydata_profiling) (3.1)\n",
            "Requirement already satisfied: tangled-up-in-unicode>=0.0.4 in /usr/local/lib/python3.10/dist-packages (from visions[type_image_path]==0.7.5->ydata_profiling) (0.2.0)\n",
            "Requirement already satisfied: graphviz in /usr/local/lib/python3.10/dist-packages (from catboost) (0.20.1)\n",
            "Requirement already satisfied: plotly in /usr/local/lib/python3.10/dist-packages (from catboost) (5.13.1)\n",
            "Requirement already satisfied: six in /usr/local/lib/python3.10/dist-packages (from catboost) (1.16.0)\n",
            "Requirement already satisfied: alembic>=1.5.0 in /usr/local/lib/python3.10/dist-packages (from optuna) (1.11.1)\n",
            "Requirement already satisfied: cmaes>=0.9.1 in /usr/local/lib/python3.10/dist-packages (from optuna) (0.10.0)\n",
            "Requirement already satisfied: colorlog in /usr/local/lib/python3.10/dist-packages (from optuna) (6.7.0)\n",
            "Requirement already satisfied: packaging>=20.0 in /usr/local/lib/python3.10/dist-packages (from optuna) (23.1)\n",
            "Requirement already satisfied: sqlalchemy>=1.3.0 in /usr/local/lib/python3.10/dist-packages (from optuna) (2.0.19)\n",
            "Requirement already satisfied: Mako in /usr/local/lib/python3.10/dist-packages (from alembic>=1.5.0->optuna) (1.2.4)\n",
            "Requirement already satisfied: typing-extensions>=4 in /usr/local/lib/python3.10/dist-packages (from alembic>=1.5.0->optuna) (4.7.1)\n",
            "Requirement already satisfied: MarkupSafe>=2.0 in /usr/local/lib/python3.10/dist-packages (from jinja2<3.2,>=2.11.1->ydata_profiling) (2.1.3)\n",
            "Requirement already satisfied: contourpy>=1.0.1 in /usr/local/lib/python3.10/dist-packages (from matplotlib>=2.2.3->phik) (1.1.0)\n",
            "Requirement already satisfied: cycler>=0.10 in /usr/local/lib/python3.10/dist-packages (from matplotlib>=2.2.3->phik) (0.11.0)\n",
            "Requirement already satisfied: fonttools>=4.22.0 in /usr/local/lib/python3.10/dist-packages (from matplotlib>=2.2.3->phik) (4.41.1)\n",
            "Requirement already satisfied: kiwisolver>=1.0.1 in /usr/local/lib/python3.10/dist-packages (from matplotlib>=2.2.3->phik) (1.4.4)\n",
            "Requirement already satisfied: pyparsing>=2.3.1 in /usr/local/lib/python3.10/dist-packages (from matplotlib>=2.2.3->phik) (3.1.0)\n",
            "Requirement already satisfied: python-dateutil>=2.7 in /usr/local/lib/python3.10/dist-packages (from matplotlib>=2.2.3->phik) (2.8.2)\n",
            "Requirement already satisfied: pytz>=2020.1 in /usr/local/lib/python3.10/dist-packages (from pandas>=0.25.1->phik) (2022.7.1)\n",
            "Requirement already satisfied: urllib3<1.27,>=1.21.1 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2.24.0->ydata_profiling) (1.26.16)\n",
            "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2.24.0->ydata_profiling) (2023.7.22)\n",
            "Requirement already satisfied: charset-normalizer~=2.0.0 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2.24.0->ydata_profiling) (2.0.12)\n",
            "Requirement already satisfied: idna<4,>=2.5 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2.24.0->ydata_profiling) (3.4)\n",
            "Requirement already satisfied: greenlet!=0.4.17 in /usr/local/lib/python3.10/dist-packages (from sqlalchemy>=1.3.0->optuna) (2.0.2)\n",
            "Requirement already satisfied: patsy>=0.5.2 in /usr/local/lib/python3.10/dist-packages (from statsmodels<1,>=0.13.2->ydata_profiling) (0.5.3)\n",
            "Requirement already satisfied: tenacity>=6.2.0 in /usr/local/lib/python3.10/dist-packages (from plotly->catboost) (8.2.2)\n"
          ]
        }
      ],
      "source": [
        "import numpy as np\n",
        "import pandas as pd\n",
        "\n",
        "from typing import Union\n",
        "\n",
        "from tqdm import tqdm\n",
        "\n",
        "import seaborn as sns\n",
        "sns.set()\n",
        "import matplotlib.pyplot as plt\n",
        "%matplotlib inline\n",
        "plt.rcParams['figure.figsize'] = (16, 9)\n",
        "plt.rcParams['axes.titlesize'] = 25\n",
        "plt.rcParams['axes.labelsize'] = 20\n",
        "\n",
        "!pip install phik ydata_profiling catboost optuna\n",
        "\n",
        "import phik\n",
        "from phik.report import plot_correlation_matrix\n",
        "\n",
        "from sklearn.model_selection import (train_test_split,\n",
        "                                     StratifiedKFold,\n",
        "                                     cross_val_score)\n",
        "from sklearn.preprocessing import (OneHotEncoder,\n",
        "                                   StandardScaler,\n",
        "                                   PolynomialFeatures)\n",
        "from sklearn.metrics import (roc_auc_score,\n",
        "                             f1_score)\n",
        "from sklearn.ensemble import RandomForestClassifier\n",
        "\n",
        "from catboost import CatBoostClassifier, cv\n",
        "from catboost.utils import Pool\n",
        "\n",
        "import optuna\n",
        "from optuna.visualization \\\n",
        "import (plot_optimization_history,\n",
        "        plot_param_importances)\n",
        "from optuna.pruners import HyperbandPruner"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Знакомимся с данными"
      ],
      "metadata": {
        "id": "1ZIRmCA3cQ8Q"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "df = pd.read_csv('/content/train.csv')\n",
        "test = pd.read_csv('/content/test.csv')"
      ],
      "metadata": {
        "id": "ALoiD6sPEW1F"
      },
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df.nunique()[df.nunique() == 1]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "I4KEhyuZdch8",
        "outputId": "c777b890-cfb7-4e2b-e176-96f30bd1b700"
      },
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "oral    1\n",
              "dtype: int64"
            ]
          },
          "metadata": {},
          "execution_count": 3
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Есть колонка с константой, удаляем:"
      ],
      "metadata": {
        "id": "oOQ6eEZcdm5G"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "df = df.drop('oral', axis='columns')\n",
        "test = test.drop('oral', axis='columns')"
      ],
      "metadata": {
        "id": "idro_PdaEi8B"
      },
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df.columns"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "RhAGAx0_drjt",
        "outputId": "50762b05-8865-4257-989a-becb016f13df"
      },
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Index(['ID', 'age', 'height(cm)', 'weight(kg)', 'waist(cm)', 'eyesight(left)',\n",
              "       'eyesight(right)', 'hearing(left)', 'hearing(right)', 'systolic',\n",
              "       'relaxation', 'fasting blood sugar', 'Cholesterol', 'triglyceride',\n",
              "       'HDL', 'LDL', 'hemoglobin', 'Urine protein', 'serum creatinine', 'AST',\n",
              "       'ALT', 'Gtp', 'dental caries', 'tartar', 'smoking'],\n",
              "      dtype='object')"
            ]
          },
          "metadata": {},
          "execution_count": 5
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Приводим колонки к *snake_case*:"
      ],
      "metadata": {
        "id": "dGSpH20XeCca"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "df.columns = df.columns.str.lower()\n",
        "test.columns = test.columns.str.lower()"
      ],
      "metadata": {
        "id": "ib_iaEscErDI"
      },
      "execution_count": 6,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df.columns = [\n",
        "    col.replace('(', '_').replace(')', '').replace(' ', '_')\n",
        "    for col in df.columns\n",
        "]\n",
        "test.columns = [\n",
        "    col.replace('(', '_').replace(')', '').replace(' ', '_')\n",
        "    for col in test.columns\n",
        "]"
      ],
      "metadata": {
        "id": "B0StL1fxEshW"
      },
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Форматируем бинарную колонку:"
      ],
      "metadata": {
        "id": "YvBPVwuwekmN"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "df['tartar'].sample(5)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Uj3KLdvKeO8D",
        "outputId": "7a82c220-5b79-4afa-8499-67dffb2bdc8f"
      },
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "4077    N\n",
              "2557    Y\n",
              "9897    Y\n",
              "3930    Y\n",
              "689     N\n",
              "Name: tartar, dtype: object"
            ]
          },
          "metadata": {},
          "execution_count": 8
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df.tartar = (df.tartar == 'Y').astype('int')\n",
        "test.tartar = (test.tartar == 'Y').astype('int')"
      ],
      "metadata": {
        "id": "2opWKI6mEt31"
      },
      "execution_count": 9,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Создадим функцию для аппроксимации полезности фичей через корреляцию *phik*:"
      ],
      "metadata": {
        "id": "VHpMkCH0expB"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def evaluate_features(arr: list) -> pd.DataFrame:\n",
        "    '''\n",
        "    Функция принимает на вход список предикторов,\n",
        "    отдаёт датафрейм, где в ячейках — коэффициент\n",
        "    корреляции phik с целевой переменной 'churned'\n",
        "    '''\n",
        "    table = df[arr + ['smoking']]\\\n",
        "    .phik_matrix()['smoking'].to_frame()\\\n",
        "    .sort_values('smoking', ascending=False).round(2)\n",
        "    return table.iloc[1:, :]"
      ],
      "metadata": {
        "id": "Y7u-U6ZdEvmd"
      },
      "execution_count": 10,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Анализ и создание признаков"
      ],
      "metadata": {
        "id": "9RxSqiqaG_Q5"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Так как я генерирую фичи для соревнования, а не для продакшна, то собственно метрика *Recall* в этом занятии для меня важнее чем *Precision* — даже когда признак скорее всего не будет полезен я не удалял его — *CatBoost* разберётся."
      ],
      "metadata": {
        "id": "DNduIaaklEPq"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Доля курящих в датасете:"
      ],
      "metadata": {
        "id": "S8N2aT9ZqFqd"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "df.smoking.mean().round(4)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "XAFz08BvqMJg",
        "outputId": "4d0fa90b-c9b0-44b2-acf9-2de1233f8fa3"
      },
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.202"
            ]
          },
          "metadata": {},
          "execution_count": 11
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Холестерин, *HDL* и *LDL*"
      ],
      "metadata": {
        "id": "H1Yx1rN4E3sR"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "evaluate_features(['cholesterol', 'hdl', 'ldl'])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 160
        },
        "id": "mfcMnmU0iIMO",
        "outputId": "9f05397b-f4d5-483c-cd26-5410ec70e1b8"
      },
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "interval columns not set, guessing: ['cholesterol', 'hdl', 'ldl', 'smoking']\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "             smoking\n",
              "hdl             0.05\n",
              "ldl             0.02\n",
              "cholesterol     0.02"
            ],
            "text/html": [
              "\n",
              "\n",
              "  <div id=\"df-ce465e54-c982-42e0-8086-57673693b614\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>smoking</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>hdl</th>\n",
              "      <td>0.05</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>ldl</th>\n",
              "      <td>0.02</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>cholesterol</th>\n",
              "      <td>0.02</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-ce465e54-c982-42e0-8086-57673693b614')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "\n",
              "\n",
              "\n",
              "    <div id=\"df-f1bf144f-905b-4e05-8679-293ee68c34c6\">\n",
              "      <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-f1bf144f-905b-4e05-8679-293ee68c34c6')\"\n",
              "              title=\"Suggest charts.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "      </button>\n",
              "    </div>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "    background-color: #E8F0FE;\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: #1967D2;\n",
              "    height: 32px;\n",
              "    padding: 0 0 0 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: #E2EBFA;\n",
              "    box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: #174EA6;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "    background-color: #3B4455;\n",
              "    fill: #D2E3FC;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart:hover {\n",
              "    background-color: #434B5C;\n",
              "    box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "    filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "    fill: #FFFFFF;\n",
              "  }\n",
              "</style>\n",
              "\n",
              "    <script>\n",
              "      async function quickchart(key) {\n",
              "        const containerElement = document.querySelector('#' + key);\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      }\n",
              "    </script>\n",
              "\n",
              "      <script>\n",
              "\n",
              "function displayQuickchartButton(domScope) {\n",
              "  let quickchartButtonEl =\n",
              "    domScope.querySelector('#df-f1bf144f-905b-4e05-8679-293ee68c34c6 button.colab-df-quickchart');\n",
              "  quickchartButtonEl.style.display =\n",
              "    google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "}\n",
              "\n",
              "        displayQuickchartButton(document);\n",
              "      </script>\n",
              "      <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-ce465e54-c982-42e0-8086-57673693b614 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-ce465e54-c982-42e0-8086-57673693b614');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 12
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Создаю три признака на основе отношений этих трёх величин.\n",
        "1. `chol_hdl`: отношение общего холестерина к *HDL*\n",
        "2. `ldl_hdl`: отношение *LDL* к *HDL*\n",
        "3. `chol_hdl_cat`: разбиваю первый на три примерные клинические категории"
      ],
      "metadata": {
        "id": "zo_o7kE-iP9p"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "df['chol_hdl'] = df['cholesterol'] / df['hdl']\n",
        "df['ldl_hdl'] = df['ldl'] / df['hdl']\n",
        "df['chol_hdl_cat'] = df['chol_hdl'].apply(lambda x:\n",
        "                                          'good' if x < 3.5\n",
        "                                          else 'okay' if x < 5\n",
        "                                          else 'bad')\n"
      ],
      "metadata": {
        "id": "rTWGg9ONE6c6"
      },
      "execution_count": 13,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "evaluate_features(['cholesterol', 'hdl', 'ldl',\n",
        "                   'chol_hdl', 'ldl_hdl', 'chol_hdl_cat'])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 254
        },
        "id": "uSSKjBMcjf4s",
        "outputId": "dadc4b85-bd4a-4202-9176-a8440d46f9bc"
      },
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "interval columns not set, guessing: ['cholesterol', 'hdl', 'ldl', 'chol_hdl', 'ldl_hdl', 'smoking']\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "              smoking\n",
              "hdl              0.05\n",
              "chol_hdl_cat     0.03\n",
              "ldl              0.02\n",
              "chol_hdl         0.02\n",
              "ldl_hdl          0.02\n",
              "cholesterol      0.02"
            ],
            "text/html": [
              "\n",
              "\n",
              "  <div id=\"df-e22f2d41-620c-4a81-b1ed-be996c784ef6\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>smoking</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>hdl</th>\n",
              "      <td>0.05</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>chol_hdl_cat</th>\n",
              "      <td>0.03</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>ldl</th>\n",
              "      <td>0.02</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>chol_hdl</th>\n",
              "      <td>0.02</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>ldl_hdl</th>\n",
              "      <td>0.02</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>cholesterol</th>\n",
              "      <td>0.02</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-e22f2d41-620c-4a81-b1ed-be996c784ef6')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "\n",
              "\n",
              "\n",
              "    <div id=\"df-641feab8-4c8d-49fe-b378-01921cbe9854\">\n",
              "      <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-641feab8-4c8d-49fe-b378-01921cbe9854')\"\n",
              "              title=\"Suggest charts.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "      </button>\n",
              "    </div>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "    background-color: #E8F0FE;\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: #1967D2;\n",
              "    height: 32px;\n",
              "    padding: 0 0 0 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: #E2EBFA;\n",
              "    box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: #174EA6;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "    background-color: #3B4455;\n",
              "    fill: #D2E3FC;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart:hover {\n",
              "    background-color: #434B5C;\n",
              "    box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "    filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "    fill: #FFFFFF;\n",
              "  }\n",
              "</style>\n",
              "\n",
              "    <script>\n",
              "      async function quickchart(key) {\n",
              "        const containerElement = document.querySelector('#' + key);\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      }\n",
              "    </script>\n",
              "\n",
              "      <script>\n",
              "\n",
              "function displayQuickchartButton(domScope) {\n",
              "  let quickchartButtonEl =\n",
              "    domScope.querySelector('#df-641feab8-4c8d-49fe-b378-01921cbe9854 button.colab-df-quickchart');\n",
              "  quickchartButtonEl.style.display =\n",
              "    google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "}\n",
              "\n",
              "        displayQuickchartButton(document);\n",
              "      </script>\n",
              "      <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-e22f2d41-620c-4a81-b1ed-be996c784ef6 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-e22f2d41-620c-4a81-b1ed-be996c784ef6');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 14
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df.groupby('chol_hdl_cat')['smoking'].agg(['count', 'mean'])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 174
        },
        "id": "KHooPorkE8pc",
        "outputId": "26d152ed-983a-4707-f8d6-00a361e5d46d"
      },
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "              count      mean\n",
              "chol_hdl_cat                 \n",
              "bad            1660  0.256627\n",
              "good           6111  0.187858\n",
              "okay           6092  0.201248"
            ],
            "text/html": [
              "\n",
              "\n",
              "  <div id=\"df-ce95bdac-7be8-41c7-a31c-250db8e9b03d\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>count</th>\n",
              "      <th>mean</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>chol_hdl_cat</th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>bad</th>\n",
              "      <td>1660</td>\n",
              "      <td>0.256627</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>good</th>\n",
              "      <td>6111</td>\n",
              "      <td>0.187858</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>okay</th>\n",
              "      <td>6092</td>\n",
              "      <td>0.201248</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-ce95bdac-7be8-41c7-a31c-250db8e9b03d')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "\n",
              "\n",
              "\n",
              "    <div id=\"df-ff878f28-2fae-4fa4-9aaf-9ae232cc54e2\">\n",
              "      <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-ff878f28-2fae-4fa4-9aaf-9ae232cc54e2')\"\n",
              "              title=\"Suggest charts.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "      </button>\n",
              "    </div>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "    background-color: #E8F0FE;\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: #1967D2;\n",
              "    height: 32px;\n",
              "    padding: 0 0 0 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: #E2EBFA;\n",
              "    box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: #174EA6;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "    background-color: #3B4455;\n",
              "    fill: #D2E3FC;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart:hover {\n",
              "    background-color: #434B5C;\n",
              "    box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "    filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "    fill: #FFFFFF;\n",
              "  }\n",
              "</style>\n",
              "\n",
              "    <script>\n",
              "      async function quickchart(key) {\n",
              "        const containerElement = document.querySelector('#' + key);\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      }\n",
              "    </script>\n",
              "\n",
              "      <script>\n",
              "\n",
              "function displayQuickchartButton(domScope) {\n",
              "  let quickchartButtonEl =\n",
              "    domScope.querySelector('#df-ff878f28-2fae-4fa4-9aaf-9ae232cc54e2 button.colab-df-quickchart');\n",
              "  quickchartButtonEl.style.display =\n",
              "    google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "}\n",
              "\n",
              "        displayQuickchartButton(document);\n",
              "      </script>\n",
              "      <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-ce95bdac-7be8-41c7-a31c-250db8e9b03d button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-ce95bdac-7be8-41c7-a31c-250db8e9b03d');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 15
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Получилось умеренно информативно, добавляю новые признаки в тестовую выборку:"
      ],
      "metadata": {
        "id": "nxF8ohvfFC4m"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "test['chol_hdl'] = test['cholesterol'] / test['hdl']\n",
        "test['ldl_hdl'] = test['ldl'] / test['hdl']\n",
        "test['chol_hdl_cat'] = test['chol_hdl'].apply(lambda x:\n",
        "                                         'good' if x < 3.5\n",
        "                                         else 'okay' if x < 5\n",
        "                                         else 'bad')"
      ],
      "metadata": {
        "id": "FUhl1ApfkGJS"
      },
      "execution_count": 16,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## *Mean Arterial Pressure*"
      ],
      "metadata": {
        "id": "gyrtJafGkLgG"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "На врачебном сайте раздобыл такую формулу для *Mean Arterial Pressure*, позволяющее оценивать систолистеское и дистолическое давление в одной метрике:\n",
        "\n",
        "$MAP$ = $(systolic$ $blood$ $pressure + (2 * diastolic$ $blood$ $pressure))$ / $3$"
      ],
      "metadata": {
        "id": "M5ykZ2g_kTQ6"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "evaluate_features(['systolic', 'relaxation'])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 129
        },
        "id": "owO8eE5Alwq9",
        "outputId": "9664e946-59bf-4845-8f59-e1a5d38362f2"
      },
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "interval columns not set, guessing: ['systolic', 'relaxation', 'smoking']\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "            smoking\n",
              "relaxation     0.02\n",
              "systolic       0.02"
            ],
            "text/html": [
              "\n",
              "\n",
              "  <div id=\"df-1b020381-d70b-4bd0-8492-9ffbfea6fe02\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>smoking</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>relaxation</th>\n",
              "      <td>0.02</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>systolic</th>\n",
              "      <td>0.02</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-1b020381-d70b-4bd0-8492-9ffbfea6fe02')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "\n",
              "\n",
              "\n",
              "    <div id=\"df-3e10ae8f-ab46-495a-8837-0286833bd638\">\n",
              "      <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-3e10ae8f-ab46-495a-8837-0286833bd638')\"\n",
              "              title=\"Suggest charts.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "      </button>\n",
              "    </div>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "    background-color: #E8F0FE;\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: #1967D2;\n",
              "    height: 32px;\n",
              "    padding: 0 0 0 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: #E2EBFA;\n",
              "    box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: #174EA6;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "    background-color: #3B4455;\n",
              "    fill: #D2E3FC;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart:hover {\n",
              "    background-color: #434B5C;\n",
              "    box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "    filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "    fill: #FFFFFF;\n",
              "  }\n",
              "</style>\n",
              "\n",
              "    <script>\n",
              "      async function quickchart(key) {\n",
              "        const containerElement = document.querySelector('#' + key);\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      }\n",
              "    </script>\n",
              "\n",
              "      <script>\n",
              "\n",
              "function displayQuickchartButton(domScope) {\n",
              "  let quickchartButtonEl =\n",
              "    domScope.querySelector('#df-3e10ae8f-ab46-495a-8837-0286833bd638 button.colab-df-quickchart');\n",
              "  quickchartButtonEl.style.display =\n",
              "    google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "}\n",
              "\n",
              "        displayQuickchartButton(document);\n",
              "      </script>\n",
              "      <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-1b020381-d70b-4bd0-8492-9ffbfea6fe02 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-1b020381-d70b-4bd0-8492-9ffbfea6fe02');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 17
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df['map'] = (df['systolic'] + (2 * df['relaxation'])) / 3\n",
        "df['map_cat'] = \\\n",
        "df.apply(lambda x: 'norm' if x['map'] < 90\n",
        "                else 'elevated',\n",
        "        axis='columns')"
      ],
      "metadata": {
        "id": "1Lr0Cs7MFP4c"
      },
      "execution_count": 18,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "evaluate_features(['systolic', 'relaxation', 'map_cat', 'map'])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 192
        },
        "id": "j8qmqKiYFUJp",
        "outputId": "5768960c-2948-4282-b84a-a04b5a0d23b6"
      },
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "interval columns not set, guessing: ['systolic', 'relaxation', 'map', 'smoking']\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "            smoking\n",
              "map_cat        0.02\n",
              "relaxation     0.02\n",
              "systolic       0.02\n",
              "map            0.02"
            ],
            "text/html": [
              "\n",
              "\n",
              "  <div id=\"df-e563da30-5530-4470-a477-84b217d42bc6\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>smoking</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>map_cat</th>\n",
              "      <td>0.02</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>relaxation</th>\n",
              "      <td>0.02</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>systolic</th>\n",
              "      <td>0.02</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>map</th>\n",
              "      <td>0.02</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-e563da30-5530-4470-a477-84b217d42bc6')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "\n",
              "\n",
              "\n",
              "    <div id=\"df-ff4b8300-2593-4fa2-a8ef-30335098493f\">\n",
              "      <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-ff4b8300-2593-4fa2-a8ef-30335098493f')\"\n",
              "              title=\"Suggest charts.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "      </button>\n",
              "    </div>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "    background-color: #E8F0FE;\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: #1967D2;\n",
              "    height: 32px;\n",
              "    padding: 0 0 0 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: #E2EBFA;\n",
              "    box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: #174EA6;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "    background-color: #3B4455;\n",
              "    fill: #D2E3FC;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart:hover {\n",
              "    background-color: #434B5C;\n",
              "    box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "    filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "    fill: #FFFFFF;\n",
              "  }\n",
              "</style>\n",
              "\n",
              "    <script>\n",
              "      async function quickchart(key) {\n",
              "        const containerElement = document.querySelector('#' + key);\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      }\n",
              "    </script>\n",
              "\n",
              "      <script>\n",
              "\n",
              "function displayQuickchartButton(domScope) {\n",
              "  let quickchartButtonEl =\n",
              "    domScope.querySelector('#df-ff4b8300-2593-4fa2-a8ef-30335098493f button.colab-df-quickchart');\n",
              "  quickchartButtonEl.style.display =\n",
              "    google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "}\n",
              "\n",
              "        displayQuickchartButton(document);\n",
              "      </script>\n",
              "      <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-e563da30-5530-4470-a477-84b217d42bc6 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-e563da30-5530-4470-a477-84b217d42bc6');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 19
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Добавляю в тестовую выборку:"
      ],
      "metadata": {
        "id": "EXcu6aTHl7jk"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "test['map'] = (test['systolic'] + (2 * test['relaxation'])) / 3\n",
        "test['map_cat'] = \\\n",
        "test.apply(lambda x: 'norm' if x['map'] < 90\n",
        "                else 'elevated',\n",
        "        axis='columns')"
      ],
      "metadata": {
        "id": "RMiwUUunmGnW"
      },
      "execution_count": 20,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## *Dental Health*"
      ],
      "metadata": {
        "id": "eBDrpBN2mJ21"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Комбинирую две бинарные метрики про здоровье зубов — *tartar* и *dental_caries*:"
      ],
      "metadata": {
        "id": "grv17iKdmhT-"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "evaluate_features(['dental_caries', 'tartar'])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 129
        },
        "id": "SAtMAnX_mdzs",
        "outputId": "a5ec87a8-fbcf-45c4-f8ab-a5b0af4b3be9"
      },
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "interval columns not set, guessing: ['dental_caries', 'tartar', 'smoking']\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "               smoking\n",
              "tartar            0.10\n",
              "dental_caries     0.09"
            ],
            "text/html": [
              "\n",
              "\n",
              "  <div id=\"df-2d45322c-f7cb-4063-aeb5-b8f5c3ac88e5\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>smoking</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>tartar</th>\n",
              "      <td>0.10</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>dental_caries</th>\n",
              "      <td>0.09</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-2d45322c-f7cb-4063-aeb5-b8f5c3ac88e5')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "\n",
              "\n",
              "\n",
              "    <div id=\"df-5b8552d6-2586-4e64-8bc1-68e3abc12866\">\n",
              "      <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-5b8552d6-2586-4e64-8bc1-68e3abc12866')\"\n",
              "              title=\"Suggest charts.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "      </button>\n",
              "    </div>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "    background-color: #E8F0FE;\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: #1967D2;\n",
              "    height: 32px;\n",
              "    padding: 0 0 0 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: #E2EBFA;\n",
              "    box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: #174EA6;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "    background-color: #3B4455;\n",
              "    fill: #D2E3FC;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart:hover {\n",
              "    background-color: #434B5C;\n",
              "    box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "    filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "    fill: #FFFFFF;\n",
              "  }\n",
              "</style>\n",
              "\n",
              "    <script>\n",
              "      async function quickchart(key) {\n",
              "        const containerElement = document.querySelector('#' + key);\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      }\n",
              "    </script>\n",
              "\n",
              "      <script>\n",
              "\n",
              "function displayQuickchartButton(domScope) {\n",
              "  let quickchartButtonEl =\n",
              "    domScope.querySelector('#df-5b8552d6-2586-4e64-8bc1-68e3abc12866 button.colab-df-quickchart');\n",
              "  quickchartButtonEl.style.display =\n",
              "    google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "}\n",
              "\n",
              "        displayQuickchartButton(document);\n",
              "      </script>\n",
              "      <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-2d45322c-f7cb-4063-aeb5-b8f5c3ac88e5 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-2d45322c-f7cb-4063-aeb5-b8f5c3ac88e5');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 21
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df['dental_health'] = df.apply(\n",
        "    lambda x: 'good'\n",
        "    if not x['tartar'] and not x['dental_caries']\n",
        "    else 'okay' if (not x['tartar'] and x['dental_caries'])\n",
        "                     or (x['tartar'] and not x['dental_caries'])\n",
        "    else 'bad',\n",
        "    axis='columns')"
      ],
      "metadata": {
        "id": "J2R_qmZTFWSP"
      },
      "execution_count": 22,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df.groupby('dental_health')['smoking']\\\n",
        ".agg(['count', 'mean']).sort_values(by='mean', ascending=False)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 174
        },
        "id": "b9mOe5l6FYJw",
        "outputId": "ae51043e-b81b-4e28-c446-33ff2e8dea37"
      },
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "               count      mean\n",
              "dental_health                 \n",
              "bad             2114  0.257332\n",
              "okay            6229  0.215444\n",
              "good            5520  0.165580"
            ],
            "text/html": [
              "\n",
              "\n",
              "  <div id=\"df-c6f5db9e-1d9f-4b60-aa1d-454f05f60156\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>count</th>\n",
              "      <th>mean</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>dental_health</th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>bad</th>\n",
              "      <td>2114</td>\n",
              "      <td>0.257332</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>okay</th>\n",
              "      <td>6229</td>\n",
              "      <td>0.215444</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>good</th>\n",
              "      <td>5520</td>\n",
              "      <td>0.165580</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-c6f5db9e-1d9f-4b60-aa1d-454f05f60156')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "\n",
              "\n",
              "\n",
              "    <div id=\"df-91f9d83a-a4ca-4784-80b9-d9ed04d8270f\">\n",
              "      <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-91f9d83a-a4ca-4784-80b9-d9ed04d8270f')\"\n",
              "              title=\"Suggest charts.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "      </button>\n",
              "    </div>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "    background-color: #E8F0FE;\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: #1967D2;\n",
              "    height: 32px;\n",
              "    padding: 0 0 0 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: #E2EBFA;\n",
              "    box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: #174EA6;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "    background-color: #3B4455;\n",
              "    fill: #D2E3FC;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart:hover {\n",
              "    background-color: #434B5C;\n",
              "    box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "    filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "    fill: #FFFFFF;\n",
              "  }\n",
              "</style>\n",
              "\n",
              "    <script>\n",
              "      async function quickchart(key) {\n",
              "        const containerElement = document.querySelector('#' + key);\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      }\n",
              "    </script>\n",
              "\n",
              "      <script>\n",
              "\n",
              "function displayQuickchartButton(domScope) {\n",
              "  let quickchartButtonEl =\n",
              "    domScope.querySelector('#df-91f9d83a-a4ca-4784-80b9-d9ed04d8270f button.colab-df-quickchart');\n",
              "  quickchartButtonEl.style.display =\n",
              "    google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "}\n",
              "\n",
              "        displayQuickchartButton(document);\n",
              "      </script>\n",
              "      <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-c6f5db9e-1d9f-4b60-aa1d-454f05f60156 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-c6f5db9e-1d9f-4b60-aa1d-454f05f60156');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 23
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "evaluate_features(['dental_health', 'tartar', 'dental_caries'])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 160
        },
        "id": "k7Rs2IESm5eY",
        "outputId": "e7f8eb54-3f8b-4307-b9ed-3f65c347e89d"
      },
      "execution_count": 24,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "interval columns not set, guessing: ['tartar', 'dental_caries', 'smoking']\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "               smoking\n",
              "tartar            0.10\n",
              "dental_caries     0.09\n",
              "dental_health     0.05"
            ],
            "text/html": [
              "\n",
              "\n",
              "  <div id=\"df-67abf408-b69b-44ec-9c1c-14521313cdc6\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>smoking</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>tartar</th>\n",
              "      <td>0.10</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>dental_caries</th>\n",
              "      <td>0.09</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>dental_health</th>\n",
              "      <td>0.05</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-67abf408-b69b-44ec-9c1c-14521313cdc6')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "\n",
              "\n",
              "\n",
              "    <div id=\"df-8797f3f1-6d01-43ac-907e-91a4353c2d9b\">\n",
              "      <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-8797f3f1-6d01-43ac-907e-91a4353c2d9b')\"\n",
              "              title=\"Suggest charts.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "      </button>\n",
              "    </div>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "    background-color: #E8F0FE;\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: #1967D2;\n",
              "    height: 32px;\n",
              "    padding: 0 0 0 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: #E2EBFA;\n",
              "    box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: #174EA6;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "    background-color: #3B4455;\n",
              "    fill: #D2E3FC;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart:hover {\n",
              "    background-color: #434B5C;\n",
              "    box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "    filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "    fill: #FFFFFF;\n",
              "  }\n",
              "</style>\n",
              "\n",
              "    <script>\n",
              "      async function quickchart(key) {\n",
              "        const containerElement = document.querySelector('#' + key);\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      }\n",
              "    </script>\n",
              "\n",
              "      <script>\n",
              "\n",
              "function displayQuickchartButton(domScope) {\n",
              "  let quickchartButtonEl =\n",
              "    domScope.querySelector('#df-8797f3f1-6d01-43ac-907e-91a4353c2d9b button.colab-df-quickchart');\n",
              "  quickchartButtonEl.style.display =\n",
              "    google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "}\n",
              "\n",
              "        displayQuickchartButton(document);\n",
              "      </script>\n",
              "      <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-67abf408-b69b-44ec-9c1c-14521313cdc6 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-67abf408-b69b-44ec-9c1c-14521313cdc6');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 24
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "test['dental_health'] = test.apply(\n",
        "    lambda x: 'good'\n",
        "    if not x['tartar'] and not x['dental_caries']\n",
        "    else 'okay' if (not x['tartar'] and x['dental_caries'])\n",
        "                     or (x['tartar'] and not x['dental_caries'])\n",
        "    else 'bad',\n",
        "    axis='columns')"
      ],
      "metadata": {
        "id": "c5OSME2PnLxp"
      },
      "execution_count": 25,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## *Waist-to-Height Ratio*"
      ],
      "metadata": {
        "id": "9iNF_8bNndW6"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Поделил окружность талии на высоту, ещё добавил клинические категории. Для мужчин и женщин они немного отличаются, поэтому усреднил."
      ],
      "metadata": {
        "id": "K8M21mZQFZt5"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "evaluate_features(['waist_cm', 'height_cm'])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 129
        },
        "id": "Mjq53Rd_p185",
        "outputId": "7090f090-6682-4042-879b-a48e01e44f0b"
      },
      "execution_count": 26,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "interval columns not set, guessing: ['waist_cm', 'height_cm', 'smoking']\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "           smoking\n",
              "waist_cm      0.06\n",
              "height_cm     0.04"
            ],
            "text/html": [
              "\n",
              "\n",
              "  <div id=\"df-a264db82-6e44-4b9b-a41b-cd2b224593c8\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>smoking</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>waist_cm</th>\n",
              "      <td>0.06</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>height_cm</th>\n",
              "      <td>0.04</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-a264db82-6e44-4b9b-a41b-cd2b224593c8')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "\n",
              "\n",
              "\n",
              "    <div id=\"df-b725d6c8-f9dd-4c89-8aed-27ebb185ac91\">\n",
              "      <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-b725d6c8-f9dd-4c89-8aed-27ebb185ac91')\"\n",
              "              title=\"Suggest charts.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "      </button>\n",
              "    </div>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "    background-color: #E8F0FE;\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: #1967D2;\n",
              "    height: 32px;\n",
              "    padding: 0 0 0 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: #E2EBFA;\n",
              "    box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: #174EA6;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "    background-color: #3B4455;\n",
              "    fill: #D2E3FC;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart:hover {\n",
              "    background-color: #434B5C;\n",
              "    box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "    filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "    fill: #FFFFFF;\n",
              "  }\n",
              "</style>\n",
              "\n",
              "    <script>\n",
              "      async function quickchart(key) {\n",
              "        const containerElement = document.querySelector('#' + key);\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      }\n",
              "    </script>\n",
              "\n",
              "      <script>\n",
              "\n",
              "function displayQuickchartButton(domScope) {\n",
              "  let quickchartButtonEl =\n",
              "    domScope.querySelector('#df-b725d6c8-f9dd-4c89-8aed-27ebb185ac91 button.colab-df-quickchart');\n",
              "  quickchartButtonEl.style.display =\n",
              "    google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "}\n",
              "\n",
              "        displayQuickchartButton(document);\n",
              "      </script>\n",
              "      <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-a264db82-6e44-4b9b-a41b-cd2b224593c8 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-a264db82-6e44-4b9b-a41b-cd2b224593c8');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 26
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df['wh_ratio'] = df['waist_cm'] / df['height_cm']\n",
        "df['wh_cat'] = \\\n",
        "df['wh_ratio'].apply(\n",
        "    lambda x:\n",
        "    'extr_slim' if x < 0.43\n",
        "    else 'slim_healthy' if x < 0.46\n",
        "    else 'healthy' if x < 0.51\n",
        "    else 'overweight' if x < 0.56\n",
        "    else 'obese'\n",
        ")"
      ],
      "metadata": {
        "id": "5Dnn-W9yFbMr"
      },
      "execution_count": 27,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df.groupby('wh_cat')['smoking']\\\n",
        ".agg(['count', 'mean']).sort_values(by='mean', ascending=False)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 237
        },
        "id": "Civ9kV5AFcso",
        "outputId": "fdc1a3e5-e3f4-4b54-896a-11775847e2ed"
      },
      "execution_count": 28,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "              count      mean\n",
              "wh_cat                       \n",
              "extr_slim       885  0.224859\n",
              "obese          1571  0.212603\n",
              "slim_healthy   1880  0.209574\n",
              "overweight     3967  0.205445\n",
              "healthy        5560  0.190288"
            ],
            "text/html": [
              "\n",
              "\n",
              "  <div id=\"df-4555ab8b-d285-40a9-b172-139d673e75e0\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>count</th>\n",
              "      <th>mean</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>wh_cat</th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>extr_slim</th>\n",
              "      <td>885</td>\n",
              "      <td>0.224859</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>obese</th>\n",
              "      <td>1571</td>\n",
              "      <td>0.212603</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>slim_healthy</th>\n",
              "      <td>1880</td>\n",
              "      <td>0.209574</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>overweight</th>\n",
              "      <td>3967</td>\n",
              "      <td>0.205445</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>healthy</th>\n",
              "      <td>5560</td>\n",
              "      <td>0.190288</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-4555ab8b-d285-40a9-b172-139d673e75e0')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "\n",
              "\n",
              "\n",
              "    <div id=\"df-e8f49929-58b5-4ea4-9186-3564d3a399ec\">\n",
              "      <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-e8f49929-58b5-4ea4-9186-3564d3a399ec')\"\n",
              "              title=\"Suggest charts.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "      </button>\n",
              "    </div>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "    background-color: #E8F0FE;\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: #1967D2;\n",
              "    height: 32px;\n",
              "    padding: 0 0 0 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: #E2EBFA;\n",
              "    box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: #174EA6;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "    background-color: #3B4455;\n",
              "    fill: #D2E3FC;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart:hover {\n",
              "    background-color: #434B5C;\n",
              "    box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "    filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "    fill: #FFFFFF;\n",
              "  }\n",
              "</style>\n",
              "\n",
              "    <script>\n",
              "      async function quickchart(key) {\n",
              "        const containerElement = document.querySelector('#' + key);\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      }\n",
              "    </script>\n",
              "\n",
              "      <script>\n",
              "\n",
              "function displayQuickchartButton(domScope) {\n",
              "  let quickchartButtonEl =\n",
              "    domScope.querySelector('#df-e8f49929-58b5-4ea4-9186-3564d3a399ec button.colab-df-quickchart');\n",
              "  quickchartButtonEl.style.display =\n",
              "    google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "}\n",
              "\n",
              "        displayQuickchartButton(document);\n",
              "      </script>\n",
              "      <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-4555ab8b-d285-40a9-b172-139d673e75e0 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-4555ab8b-d285-40a9-b172-139d673e75e0');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 28
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "evaluate_features(['waist_cm', 'height_cm', 'wh_ratio', 'wh_cat'])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 192
        },
        "id": "Q5lSD9boFeGS",
        "outputId": "1eae0a5f-fa31-48da-bef9-af393ba4821f"
      },
      "execution_count": 29,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "interval columns not set, guessing: ['waist_cm', 'height_cm', 'wh_ratio', 'smoking']\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "           smoking\n",
              "waist_cm      0.06\n",
              "height_cm     0.04\n",
              "wh_ratio      0.03\n",
              "wh_cat        0.02"
            ],
            "text/html": [
              "\n",
              "\n",
              "  <div id=\"df-6ea1db6b-b367-45d2-8589-b1c03c1e52b3\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>smoking</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>waist_cm</th>\n",
              "      <td>0.06</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>height_cm</th>\n",
              "      <td>0.04</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>wh_ratio</th>\n",
              "      <td>0.03</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>wh_cat</th>\n",
              "      <td>0.02</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-6ea1db6b-b367-45d2-8589-b1c03c1e52b3')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "\n",
              "\n",
              "\n",
              "    <div id=\"df-7f915355-ce8d-45ca-9358-21617cc60f2c\">\n",
              "      <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-7f915355-ce8d-45ca-9358-21617cc60f2c')\"\n",
              "              title=\"Suggest charts.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "      </button>\n",
              "    </div>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "    background-color: #E8F0FE;\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: #1967D2;\n",
              "    height: 32px;\n",
              "    padding: 0 0 0 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: #E2EBFA;\n",
              "    box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: #174EA6;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "    background-color: #3B4455;\n",
              "    fill: #D2E3FC;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart:hover {\n",
              "    background-color: #434B5C;\n",
              "    box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "    filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "    fill: #FFFFFF;\n",
              "  }\n",
              "</style>\n",
              "\n",
              "    <script>\n",
              "      async function quickchart(key) {\n",
              "        const containerElement = document.querySelector('#' + key);\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      }\n",
              "    </script>\n",
              "\n",
              "      <script>\n",
              "\n",
              "function displayQuickchartButton(domScope) {\n",
              "  let quickchartButtonEl =\n",
              "    domScope.querySelector('#df-7f915355-ce8d-45ca-9358-21617cc60f2c button.colab-df-quickchart');\n",
              "  quickchartButtonEl.style.display =\n",
              "    google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "}\n",
              "\n",
              "        displayQuickchartButton(document);\n",
              "      </script>\n",
              "      <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-6ea1db6b-b367-45d2-8589-b1c03c1e52b3 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-6ea1db6b-b367-45d2-8589-b1c03c1e52b3');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 29
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "test['wh_ratio'] = test['waist_cm'] / test['height_cm']\n",
        "test['wh_cat'] = \\\n",
        "test['wh_ratio'].apply(\n",
        "    lambda x:\n",
        "    'extr_slim' if x < 0.43\n",
        "    else 'slim_healthy' if x < 0.46\n",
        "    else 'healthy' if x < 0.51\n",
        "    else 'overweight' if x < 0.56\n",
        "    else 'obese')"
      ],
      "metadata": {
        "id": "vVM1M6FVqsXe"
      },
      "execution_count": 30,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Зрение и слух"
      ],
      "metadata": {
        "id": "tNQi5ZcPq1Lh"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Логично, что стоит объединить данные по левому и правому уху, а также левому и правому глазу."
      ],
      "metadata": {
        "id": "TcqIpx_7rELF"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "df['hearing_mean'] = (df['hearing_left'] + df['hearing_right']) / 2\n",
        "df['eyesight_mean'] = (df['eyesight_left'] + df['eyesight_right']) / 2"
      ],
      "metadata": {
        "id": "hvnGcJl8roqK"
      },
      "execution_count": 31,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "evaluate_features(['hearing_mean', 'eyesight_mean',\n",
        "                   'eyesight_right', 'eyesight_left',\n",
        "                   'hearing_left', 'hearing_right'])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 274
        },
        "id": "Den0Zl6GHsF6",
        "outputId": "9f3af3f4-c3bf-434f-a22d-c2fa6e272f90"
      },
      "execution_count": 32,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "interval columns not set, guessing: ['hearing_mean', 'eyesight_mean', 'eyesight_right', 'eyesight_left', 'hearing_left', 'hearing_right', 'smoking']\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                smoking\n",
              "hearing_left       0.04\n",
              "hearing_right      0.02\n",
              "hearing_mean       0.01\n",
              "eyesight_mean      0.01\n",
              "eyesight_right     0.00\n",
              "eyesight_left      0.00"
            ],
            "text/html": [
              "\n",
              "\n",
              "  <div id=\"df-deeb843c-e94f-4b58-a0fe-16932c338b57\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>smoking</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>hearing_left</th>\n",
              "      <td>0.04</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>hearing_right</th>\n",
              "      <td>0.02</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>hearing_mean</th>\n",
              "      <td>0.01</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>eyesight_mean</th>\n",
              "      <td>0.01</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>eyesight_right</th>\n",
              "      <td>0.00</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>eyesight_left</th>\n",
              "      <td>0.00</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-deeb843c-e94f-4b58-a0fe-16932c338b57')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "\n",
              "\n",
              "\n",
              "    <div id=\"df-fe59a1b8-e0c6-4ea7-bb2f-112f1c9406be\">\n",
              "      <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-fe59a1b8-e0c6-4ea7-bb2f-112f1c9406be')\"\n",
              "              title=\"Suggest charts.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "      </button>\n",
              "    </div>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "    background-color: #E8F0FE;\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: #1967D2;\n",
              "    height: 32px;\n",
              "    padding: 0 0 0 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: #E2EBFA;\n",
              "    box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: #174EA6;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "    background-color: #3B4455;\n",
              "    fill: #D2E3FC;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart:hover {\n",
              "    background-color: #434B5C;\n",
              "    box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "    filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "    fill: #FFFFFF;\n",
              "  }\n",
              "</style>\n",
              "\n",
              "    <script>\n",
              "      async function quickchart(key) {\n",
              "        const containerElement = document.querySelector('#' + key);\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      }\n",
              "    </script>\n",
              "\n",
              "      <script>\n",
              "\n",
              "function displayQuickchartButton(domScope) {\n",
              "  let quickchartButtonEl =\n",
              "    domScope.querySelector('#df-fe59a1b8-e0c6-4ea7-bb2f-112f1c9406be button.colab-df-quickchart');\n",
              "  quickchartButtonEl.style.display =\n",
              "    google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "}\n",
              "\n",
              "        displayQuickchartButton(document);\n",
              "      </script>\n",
              "      <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-deeb843c-e94f-4b58-a0fe-16932c338b57 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-deeb843c-e94f-4b58-a0fe-16932c338b57');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 32
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "test['hearing_mean'] = (test['hearing_left'] + test['hearing_right']) / 2\n",
        "test['eyesight_mean'] = (test['eyesight_left'] + test['eyesight_right']) / 2"
      ],
      "metadata": {
        "id": "piEG5TvpLikB"
      },
      "execution_count": 33,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Возраст"
      ],
      "metadata": {
        "id": "apaWPKqRsLfv"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "evaluate_features(['age'])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 98
        },
        "id": "az23ZdFwtOIc",
        "outputId": "74ae3674-1167-4953-fc49-0099b18e0ffc"
      },
      "execution_count": 34,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "interval columns not set, guessing: ['age', 'smoking']\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "     smoking\n",
              "age     0.18"
            ],
            "text/html": [
              "\n",
              "\n",
              "  <div id=\"df-4ef14fd8-ee93-4bba-a1e5-cd21148ce330\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>smoking</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>age</th>\n",
              "      <td>0.18</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-4ef14fd8-ee93-4bba-a1e5-cd21148ce330')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "\n",
              "\n",
              "\n",
              "    <div id=\"df-945f6049-46c0-4345-a583-e0ffeb4df162\">\n",
              "      <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-945f6049-46c0-4345-a583-e0ffeb4df162')\"\n",
              "              title=\"Suggest charts.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "      </button>\n",
              "    </div>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "    background-color: #E8F0FE;\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: #1967D2;\n",
              "    height: 32px;\n",
              "    padding: 0 0 0 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: #E2EBFA;\n",
              "    box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: #174EA6;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "    background-color: #3B4455;\n",
              "    fill: #D2E3FC;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart:hover {\n",
              "    background-color: #434B5C;\n",
              "    box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "    filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "    fill: #FFFFFF;\n",
              "  }\n",
              "</style>\n",
              "\n",
              "    <script>\n",
              "      async function quickchart(key) {\n",
              "        const containerElement = document.querySelector('#' + key);\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      }\n",
              "    </script>\n",
              "\n",
              "      <script>\n",
              "\n",
              "function displayQuickchartButton(domScope) {\n",
              "  let quickchartButtonEl =\n",
              "    domScope.querySelector('#df-945f6049-46c0-4345-a583-e0ffeb4df162 button.colab-df-quickchart');\n",
              "  quickchartButtonEl.style.display =\n",
              "    google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "}\n",
              "\n",
              "        displayQuickchartButton(document);\n",
              "      </script>\n",
              "      <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-4ef14fd8-ee93-4bba-a1e5-cd21148ce330 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-4ef14fd8-ee93-4bba-a1e5-cd21148ce330');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 34
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df.groupby('age')['smoking'].agg(['count', 'mean'])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 519
        },
        "id": "RPbXGpV4sb-l",
        "outputId": "d81e69d4-6a7b-4490-ae8a-c951fb631e1d"
      },
      "execution_count": 35,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "     count      mean\n",
              "age                 \n",
              "20     654  0.171254\n",
              "25    1558  0.165597\n",
              "30    1711  0.163647\n",
              "35    1765  0.171671\n",
              "40    2662  0.276108\n",
              "45    1235  0.273684\n",
              "50     995  0.257286\n",
              "55     988  0.210526\n",
              "60    1369  0.168736\n",
              "65     365  0.126027\n",
              "70     225  0.071111\n",
              "75     224  0.053571\n",
              "80     104  0.048077\n",
              "85       8  0.000000"
            ],
            "text/html": [
              "\n",
              "\n",
              "  <div id=\"df-89074513-8fe7-4821-9993-2c5f43c3947e\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>count</th>\n",
              "      <th>mean</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>age</th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>20</th>\n",
              "      <td>654</td>\n",
              "      <td>0.171254</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>25</th>\n",
              "      <td>1558</td>\n",
              "      <td>0.165597</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>30</th>\n",
              "      <td>1711</td>\n",
              "      <td>0.163647</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>35</th>\n",
              "      <td>1765</td>\n",
              "      <td>0.171671</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>40</th>\n",
              "      <td>2662</td>\n",
              "      <td>0.276108</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>45</th>\n",
              "      <td>1235</td>\n",
              "      <td>0.273684</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>50</th>\n",
              "      <td>995</td>\n",
              "      <td>0.257286</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>55</th>\n",
              "      <td>988</td>\n",
              "      <td>0.210526</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>60</th>\n",
              "      <td>1369</td>\n",
              "      <td>0.168736</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>65</th>\n",
              "      <td>365</td>\n",
              "      <td>0.126027</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>70</th>\n",
              "      <td>225</td>\n",
              "      <td>0.071111</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>75</th>\n",
              "      <td>224</td>\n",
              "      <td>0.053571</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>80</th>\n",
              "      <td>104</td>\n",
              "      <td>0.048077</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>85</th>\n",
              "      <td>8</td>\n",
              "      <td>0.000000</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-89074513-8fe7-4821-9993-2c5f43c3947e')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "\n",
              "\n",
              "\n",
              "    <div id=\"df-8e4bffd9-7fa7-497f-91b9-26d731b8fb92\">\n",
              "      <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-8e4bffd9-7fa7-497f-91b9-26d731b8fb92')\"\n",
              "              title=\"Suggest charts.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "      </button>\n",
              "    </div>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "    background-color: #E8F0FE;\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: #1967D2;\n",
              "    height: 32px;\n",
              "    padding: 0 0 0 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: #E2EBFA;\n",
              "    box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: #174EA6;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "    background-color: #3B4455;\n",
              "    fill: #D2E3FC;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart:hover {\n",
              "    background-color: #434B5C;\n",
              "    box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "    filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "    fill: #FFFFFF;\n",
              "  }\n",
              "</style>\n",
              "\n",
              "    <script>\n",
              "      async function quickchart(key) {\n",
              "        const containerElement = document.querySelector('#' + key);\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      }\n",
              "    </script>\n",
              "\n",
              "      <script>\n",
              "\n",
              "function displayQuickchartButton(domScope) {\n",
              "  let quickchartButtonEl =\n",
              "    domScope.querySelector('#df-8e4bffd9-7fa7-497f-91b9-26d731b8fb92 button.colab-df-quickchart');\n",
              "  quickchartButtonEl.style.display =\n",
              "    google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "}\n",
              "\n",
              "        displayQuickchartButton(document);\n",
              "      </script>\n",
              "      <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-89074513-8fe7-4821-9993-2c5f43c3947e button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-89074513-8fe7-4821-9993-2c5f43c3947e');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 35
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Хоть алгоритм градиентного бустинга и может восстанавливать нелинейные связи самостоятельно, можно ему в этом помочь поделив непрерывный признак на информативные \"корзины\" (*bins*)"
      ],
      "metadata": {
        "id": "c5DUtrQis_29"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "df['middle_aged'] = (df['age'] >= 40) & (df['age'] <= 55)\n",
        "df['senior'] = df['age'] >= 65"
      ],
      "metadata": {
        "id": "p9lp4wS5tVTR"
      },
      "execution_count": 36,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "evaluate_features(['middle_aged', 'senior', 'age'])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 160
        },
        "id": "pQHGuIvvtXlh",
        "outputId": "deb15071-32fd-418f-9d3a-09f3dd218454"
      },
      "execution_count": 37,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "interval columns not set, guessing: ['age', 'smoking']\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "             smoking\n",
              "middle_aged     0.20\n",
              "age             0.18\n",
              "senior          0.12"
            ],
            "text/html": [
              "\n",
              "\n",
              "  <div id=\"df-9dd6a95e-8390-40c5-82e4-7961a49bbbf2\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>smoking</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>middle_aged</th>\n",
              "      <td>0.20</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>age</th>\n",
              "      <td>0.18</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>senior</th>\n",
              "      <td>0.12</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-9dd6a95e-8390-40c5-82e4-7961a49bbbf2')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "\n",
              "\n",
              "\n",
              "    <div id=\"df-80de9c0f-838a-43ae-940d-66343c13ec99\">\n",
              "      <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-80de9c0f-838a-43ae-940d-66343c13ec99')\"\n",
              "              title=\"Suggest charts.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "      </button>\n",
              "    </div>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "    background-color: #E8F0FE;\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: #1967D2;\n",
              "    height: 32px;\n",
              "    padding: 0 0 0 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: #E2EBFA;\n",
              "    box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: #174EA6;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "    background-color: #3B4455;\n",
              "    fill: #D2E3FC;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart:hover {\n",
              "    background-color: #434B5C;\n",
              "    box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "    filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "    fill: #FFFFFF;\n",
              "  }\n",
              "</style>\n",
              "\n",
              "    <script>\n",
              "      async function quickchart(key) {\n",
              "        const containerElement = document.querySelector('#' + key);\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      }\n",
              "    </script>\n",
              "\n",
              "      <script>\n",
              "\n",
              "function displayQuickchartButton(domScope) {\n",
              "  let quickchartButtonEl =\n",
              "    domScope.querySelector('#df-80de9c0f-838a-43ae-940d-66343c13ec99 button.colab-df-quickchart');\n",
              "  quickchartButtonEl.style.display =\n",
              "    google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "}\n",
              "\n",
              "        displayQuickchartButton(document);\n",
              "      </script>\n",
              "      <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-9dd6a95e-8390-40c5-82e4-7961a49bbbf2 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-9dd6a95e-8390-40c5-82e4-7961a49bbbf2');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 37
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "test['middle_aged'] = (test['age'] >= 40) & (test['age'] <= 55)\n",
        "test['senior'] = test['age'] >= 65"
      ],
      "metadata": {
        "id": "AMFPpfQ1tldZ"
      },
      "execution_count": 38,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Вес, индекс массы тела"
      ],
      "metadata": {
        "id": "BXvI5X3PtmlK"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "evaluate_features(['weight_kg'])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 98
        },
        "id": "WpAaNF1ixjS5",
        "outputId": "1d7b6e46-0866-41cd-ff2e-85b14a77f6d4"
      },
      "execution_count": 39,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "interval columns not set, guessing: ['weight_kg', 'smoking']\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "           smoking\n",
              "weight_kg     0.04"
            ],
            "text/html": [
              "\n",
              "\n",
              "  <div id=\"df-48a07c15-7997-4a33-8427-5098f1ac73c4\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>smoking</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>weight_kg</th>\n",
              "      <td>0.04</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-48a07c15-7997-4a33-8427-5098f1ac73c4')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "\n",
              "\n",
              "\n",
              "    <div id=\"df-7db8d369-dc4a-4bb7-a76d-0b3bb01d4fab\">\n",
              "      <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-7db8d369-dc4a-4bb7-a76d-0b3bb01d4fab')\"\n",
              "              title=\"Suggest charts.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "      </button>\n",
              "    </div>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "    background-color: #E8F0FE;\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: #1967D2;\n",
              "    height: 32px;\n",
              "    padding: 0 0 0 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: #E2EBFA;\n",
              "    box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: #174EA6;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "    background-color: #3B4455;\n",
              "    fill: #D2E3FC;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart:hover {\n",
              "    background-color: #434B5C;\n",
              "    box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "    filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "    fill: #FFFFFF;\n",
              "  }\n",
              "</style>\n",
              "\n",
              "    <script>\n",
              "      async function quickchart(key) {\n",
              "        const containerElement = document.querySelector('#' + key);\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      }\n",
              "    </script>\n",
              "\n",
              "      <script>\n",
              "\n",
              "function displayQuickchartButton(domScope) {\n",
              "  let quickchartButtonEl =\n",
              "    domScope.querySelector('#df-7db8d369-dc4a-4bb7-a76d-0b3bb01d4fab button.colab-df-quickchart');\n",
              "  quickchartButtonEl.style.display =\n",
              "    google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "}\n",
              "\n",
              "        displayQuickchartButton(document);\n",
              "      </script>\n",
              "      <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-48a07c15-7997-4a33-8427-5098f1ac73c4 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-48a07c15-7997-4a33-8427-5098f1ac73c4');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 39
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df.groupby('weight_kg')['smoking'].agg(['count', 'mean'])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 676
        },
        "id": "zku610N8xm13",
        "outputId": "cf70be42-6d29-49d8-f03b-a6ce96797027"
      },
      "execution_count": 40,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "           count      mean\n",
              "weight_kg                 \n",
              "40             5  0.200000\n",
              "45            60  0.200000\n",
              "50           282  0.297872\n",
              "55           840  0.203571\n",
              "60          1754  0.185291\n",
              "65          2549  0.205571\n",
              "70          2662  0.185950\n",
              "75          2212  0.188517\n",
              "80          1481  0.206617\n",
              "85           880  0.231818\n",
              "90           520  0.228846\n",
              "95           313  0.239617\n",
              "100          162  0.246914\n",
              "105           67  0.194030\n",
              "110           48  0.208333\n",
              "115           16  0.125000\n",
              "120            6  0.000000\n",
              "125            3  0.333333\n",
              "130            3  0.333333"
            ],
            "text/html": [
              "\n",
              "\n",
              "  <div id=\"df-1493a746-66a2-4535-982b-3691f55db090\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>count</th>\n",
              "      <th>mean</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>weight_kg</th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>40</th>\n",
              "      <td>5</td>\n",
              "      <td>0.200000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>45</th>\n",
              "      <td>60</td>\n",
              "      <td>0.200000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>50</th>\n",
              "      <td>282</td>\n",
              "      <td>0.297872</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>55</th>\n",
              "      <td>840</td>\n",
              "      <td>0.203571</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>60</th>\n",
              "      <td>1754</td>\n",
              "      <td>0.185291</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>65</th>\n",
              "      <td>2549</td>\n",
              "      <td>0.205571</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>70</th>\n",
              "      <td>2662</td>\n",
              "      <td>0.185950</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>75</th>\n",
              "      <td>2212</td>\n",
              "      <td>0.188517</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>80</th>\n",
              "      <td>1481</td>\n",
              "      <td>0.206617</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>85</th>\n",
              "      <td>880</td>\n",
              "      <td>0.231818</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>90</th>\n",
              "      <td>520</td>\n",
              "      <td>0.228846</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>95</th>\n",
              "      <td>313</td>\n",
              "      <td>0.239617</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>100</th>\n",
              "      <td>162</td>\n",
              "      <td>0.246914</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>105</th>\n",
              "      <td>67</td>\n",
              "      <td>0.194030</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>110</th>\n",
              "      <td>48</td>\n",
              "      <td>0.208333</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>115</th>\n",
              "      <td>16</td>\n",
              "      <td>0.125000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>120</th>\n",
              "      <td>6</td>\n",
              "      <td>0.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>125</th>\n",
              "      <td>3</td>\n",
              "      <td>0.333333</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>130</th>\n",
              "      <td>3</td>\n",
              "      <td>0.333333</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-1493a746-66a2-4535-982b-3691f55db090')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "\n",
              "\n",
              "\n",
              "    <div id=\"df-086d9d91-cfcc-4568-a360-6d0ba7338842\">\n",
              "      <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-086d9d91-cfcc-4568-a360-6d0ba7338842')\"\n",
              "              title=\"Suggest charts.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "      </button>\n",
              "    </div>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "    background-color: #E8F0FE;\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: #1967D2;\n",
              "    height: 32px;\n",
              "    padding: 0 0 0 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: #E2EBFA;\n",
              "    box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: #174EA6;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "    background-color: #3B4455;\n",
              "    fill: #D2E3FC;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart:hover {\n",
              "    background-color: #434B5C;\n",
              "    box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "    filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "    fill: #FFFFFF;\n",
              "  }\n",
              "</style>\n",
              "\n",
              "    <script>\n",
              "      async function quickchart(key) {\n",
              "        const containerElement = document.querySelector('#' + key);\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      }\n",
              "    </script>\n",
              "\n",
              "      <script>\n",
              "\n",
              "function displayQuickchartButton(domScope) {\n",
              "  let quickchartButtonEl =\n",
              "    domScope.querySelector('#df-086d9d91-cfcc-4568-a360-6d0ba7338842 button.colab-df-quickchart');\n",
              "  quickchartButtonEl.style.display =\n",
              "    google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "}\n",
              "\n",
              "        displayQuickchartButton(document);\n",
              "      </script>\n",
              "      <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-1493a746-66a2-4535-982b-3691f55db090 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-1493a746-66a2-4535-982b-3691f55db090');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 40
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df['normal_weight'] = (df['weight_kg'] >= 55) & (df['weight_kg'] <= 80)"
      ],
      "metadata": {
        "id": "alcmawG4xhnk"
      },
      "execution_count": 41,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "evaluate_features(['weight_kg', 'normal_weight'])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 129
        },
        "id": "R2zdhRkSx7Fv",
        "outputId": "513ab91c-0379-4bf3-866b-388be7de0356"
      },
      "execution_count": 42,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "interval columns not set, guessing: ['weight_kg', 'smoking']\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "               smoking\n",
              "normal_weight     0.06\n",
              "weight_kg         0.04"
            ],
            "text/html": [
              "\n",
              "\n",
              "  <div id=\"df-16009d30-3891-42a5-8670-9cabe04b94f1\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>smoking</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>normal_weight</th>\n",
              "      <td>0.06</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>weight_kg</th>\n",
              "      <td>0.04</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-16009d30-3891-42a5-8670-9cabe04b94f1')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "\n",
              "\n",
              "\n",
              "    <div id=\"df-9e90968f-9f31-4c9a-b8a2-feed95cdcaac\">\n",
              "      <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-9e90968f-9f31-4c9a-b8a2-feed95cdcaac')\"\n",
              "              title=\"Suggest charts.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "      </button>\n",
              "    </div>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "    background-color: #E8F0FE;\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: #1967D2;\n",
              "    height: 32px;\n",
              "    padding: 0 0 0 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: #E2EBFA;\n",
              "    box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: #174EA6;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "    background-color: #3B4455;\n",
              "    fill: #D2E3FC;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart:hover {\n",
              "    background-color: #434B5C;\n",
              "    box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "    filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "    fill: #FFFFFF;\n",
              "  }\n",
              "</style>\n",
              "\n",
              "    <script>\n",
              "      async function quickchart(key) {\n",
              "        const containerElement = document.querySelector('#' + key);\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      }\n",
              "    </script>\n",
              "\n",
              "      <script>\n",
              "\n",
              "function displayQuickchartButton(domScope) {\n",
              "  let quickchartButtonEl =\n",
              "    domScope.querySelector('#df-9e90968f-9f31-4c9a-b8a2-feed95cdcaac button.colab-df-quickchart');\n",
              "  quickchartButtonEl.style.display =\n",
              "    google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "}\n",
              "\n",
              "        displayQuickchartButton(document);\n",
              "      </script>\n",
              "      <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-16009d30-3891-42a5-8670-9cabe04b94f1 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-16009d30-3891-42a5-8670-9cabe04b94f1');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 42
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "test['normal_weight'] = (test['weight_kg'] >= 55) & (test['weight_kg'] <= 80)"
      ],
      "metadata": {
        "id": "qlnHoC1TyAXH"
      },
      "execution_count": 43,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Считаю индекс массы тела (*BMI*) по формуле\n",
        "\n",
        "$BMI = weight / height^2$\n",
        "\n",
        "Создаю ещё один признак по клиническим категориям:"
      ],
      "metadata": {
        "id": "sXgZNv_RyPO-"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "df['bmi'] = df['weight_kg'] / ((df['height_cm'] * 0.01) ** 2)\n",
        "df['bmi_cat'] = df.bmi.apply(lambda x: 'underweight' if x < 18.5\n",
        "                             else 'healthy' if x >= 18.5 and x < 25\n",
        "                             else 'overweight' if x >= 25 and x < 30\n",
        "                             else 'obese')"
      ],
      "metadata": {
        "id": "R7mm9IV8yJzC"
      },
      "execution_count": 44,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "evaluate_features(['weight_kg', 'normal_weight', 'bmi', 'bmi_cat'])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 192
        },
        "id": "wUzcde3nyj7N",
        "outputId": "0a768c16-cc6c-47fc-ff65-811bc747354c"
      },
      "execution_count": 45,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "interval columns not set, guessing: ['weight_kg', 'bmi', 'smoking']\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "               smoking\n",
              "bmi               0.07\n",
              "normal_weight     0.06\n",
              "bmi_cat           0.05\n",
              "weight_kg         0.04"
            ],
            "text/html": [
              "\n",
              "\n",
              "  <div id=\"df-c3454cbf-ead5-4958-ad0a-752fb0f1a127\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>smoking</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>bmi</th>\n",
              "      <td>0.07</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>normal_weight</th>\n",
              "      <td>0.06</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>bmi_cat</th>\n",
              "      <td>0.05</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>weight_kg</th>\n",
              "      <td>0.04</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-c3454cbf-ead5-4958-ad0a-752fb0f1a127')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "\n",
              "\n",
              "\n",
              "    <div id=\"df-81b59614-e875-4da5-b046-cb6786fd0ad9\">\n",
              "      <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-81b59614-e875-4da5-b046-cb6786fd0ad9')\"\n",
              "              title=\"Suggest charts.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "      </button>\n",
              "    </div>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "    background-color: #E8F0FE;\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: #1967D2;\n",
              "    height: 32px;\n",
              "    padding: 0 0 0 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: #E2EBFA;\n",
              "    box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: #174EA6;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "    background-color: #3B4455;\n",
              "    fill: #D2E3FC;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart:hover {\n",
              "    background-color: #434B5C;\n",
              "    box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "    filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "    fill: #FFFFFF;\n",
              "  }\n",
              "</style>\n",
              "\n",
              "    <script>\n",
              "      async function quickchart(key) {\n",
              "        const containerElement = document.querySelector('#' + key);\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      }\n",
              "    </script>\n",
              "\n",
              "      <script>\n",
              "\n",
              "function displayQuickchartButton(domScope) {\n",
              "  let quickchartButtonEl =\n",
              "    domScope.querySelector('#df-81b59614-e875-4da5-b046-cb6786fd0ad9 button.colab-df-quickchart');\n",
              "  quickchartButtonEl.style.display =\n",
              "    google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "}\n",
              "\n",
              "        displayQuickchartButton(document);\n",
              "      </script>\n",
              "      <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-c3454cbf-ead5-4958-ad0a-752fb0f1a127 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-c3454cbf-ead5-4958-ad0a-752fb0f1a127');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 45
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "test['bmi'] = test['weight_kg'] / ((test['height_cm'] * 0.01) ** 2)\n",
        "test['bmi_cat'] = test.bmi.apply(lambda x: 'underweight' if x < 18.5\n",
        "                             else 'healthy' if x >= 18.5 and x < 25\n",
        "                             else 'overweight' if x >= 25 and x < 30\n",
        "                             else 'obese')"
      ],
      "metadata": {
        "id": "shAl_NQOyw2_"
      },
      "execution_count": 46,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Показатели работы печени"
      ],
      "metadata": {
        "id": "6Jz3loc53BdT"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Отношение АСТ к АЛТ называется [Коэффициентом де Ритиса](https://ru.wikipedia.org/wiki/%D0%9A%D0%BE%D1%8D%D1%84%D1%84%D0%B8%D1%86%D0%B8%D0%B5%D0%BD%D1%82_%D0%B4%D0%B5_%D0%A0%D0%B8%D1%82%D0%B8%D1%81%D0%B0)."
      ],
      "metadata": {
        "id": "YYKet9td3iyP"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "evaluate_features(['ast', 'alt'])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 129
        },
        "id": "7zHbGeiT30-g",
        "outputId": "304248c7-c31c-4fea-b7b9-31c08c2ce27a"
      },
      "execution_count": 47,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "interval columns not set, guessing: ['ast', 'alt', 'smoking']\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "     smoking\n",
              "alt     0.01\n",
              "ast     0.00"
            ],
            "text/html": [
              "\n",
              "\n",
              "  <div id=\"df-b3a84151-5ec7-497f-9bdf-2fb81f0b5bbe\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>smoking</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>alt</th>\n",
              "      <td>0.01</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>ast</th>\n",
              "      <td>0.00</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-b3a84151-5ec7-497f-9bdf-2fb81f0b5bbe')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "\n",
              "\n",
              "\n",
              "    <div id=\"df-680035ed-ecb6-4ea0-8323-0b3892c9da2c\">\n",
              "      <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-680035ed-ecb6-4ea0-8323-0b3892c9da2c')\"\n",
              "              title=\"Suggest charts.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "      </button>\n",
              "    </div>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "    background-color: #E8F0FE;\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: #1967D2;\n",
              "    height: 32px;\n",
              "    padding: 0 0 0 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: #E2EBFA;\n",
              "    box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: #174EA6;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "    background-color: #3B4455;\n",
              "    fill: #D2E3FC;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart:hover {\n",
              "    background-color: #434B5C;\n",
              "    box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "    filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "    fill: #FFFFFF;\n",
              "  }\n",
              "</style>\n",
              "\n",
              "    <script>\n",
              "      async function quickchart(key) {\n",
              "        const containerElement = document.querySelector('#' + key);\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      }\n",
              "    </script>\n",
              "\n",
              "      <script>\n",
              "\n",
              "function displayQuickchartButton(domScope) {\n",
              "  let quickchartButtonEl =\n",
              "    domScope.querySelector('#df-680035ed-ecb6-4ea0-8323-0b3892c9da2c button.colab-df-quickchart');\n",
              "  quickchartButtonEl.style.display =\n",
              "    google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "}\n",
              "\n",
              "        displayQuickchartButton(document);\n",
              "      </script>\n",
              "      <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-b3a84151-5ec7-497f-9bdf-2fb81f0b5bbe button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-b3a84151-5ec7-497f-9bdf-2fb81f0b5bbe');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 47
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df['ast_alt'] = df['ast'] / df['alt']"
      ],
      "metadata": {
        "id": "JxX5Jr0v34ku"
      },
      "execution_count": 48,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "evaluate_features(['ast', 'alt', 'ast_alt'])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 160
        },
        "id": "lvYl-lp63702",
        "outputId": "22b36dfa-4806-419b-ac4a-f7afcfb38be7"
      },
      "execution_count": 49,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "interval columns not set, guessing: ['ast', 'alt', 'ast_alt', 'smoking']\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "         smoking\n",
              "ast_alt     0.03\n",
              "alt         0.01\n",
              "ast         0.00"
            ],
            "text/html": [
              "\n",
              "\n",
              "  <div id=\"df-58505b51-faf6-4f82-b21c-ad26d4336e9d\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>smoking</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>ast_alt</th>\n",
              "      <td>0.03</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>alt</th>\n",
              "      <td>0.01</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>ast</th>\n",
              "      <td>0.00</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-58505b51-faf6-4f82-b21c-ad26d4336e9d')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "\n",
              "\n",
              "\n",
              "    <div id=\"df-01d513fd-d245-4c09-ad27-16447d4deffb\">\n",
              "      <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-01d513fd-d245-4c09-ad27-16447d4deffb')\"\n",
              "              title=\"Suggest charts.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "      </button>\n",
              "    </div>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "    background-color: #E8F0FE;\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: #1967D2;\n",
              "    height: 32px;\n",
              "    padding: 0 0 0 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: #E2EBFA;\n",
              "    box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: #174EA6;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "    background-color: #3B4455;\n",
              "    fill: #D2E3FC;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart:hover {\n",
              "    background-color: #434B5C;\n",
              "    box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "    filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "    fill: #FFFFFF;\n",
              "  }\n",
              "</style>\n",
              "\n",
              "    <script>\n",
              "      async function quickchart(key) {\n",
              "        const containerElement = document.querySelector('#' + key);\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      }\n",
              "    </script>\n",
              "\n",
              "      <script>\n",
              "\n",
              "function displayQuickchartButton(domScope) {\n",
              "  let quickchartButtonEl =\n",
              "    domScope.querySelector('#df-01d513fd-d245-4c09-ad27-16447d4deffb button.colab-df-quickchart');\n",
              "  quickchartButtonEl.style.display =\n",
              "    google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "}\n",
              "\n",
              "        displayQuickchartButton(document);\n",
              "      </script>\n",
              "      <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-58505b51-faf6-4f82-b21c-ad26d4336e9d button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-58505b51-faf6-4f82-b21c-ad26d4336e9d');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 49
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "test['ast_alt'] = test['ast'] / test['alt']"
      ],
      "metadata": {
        "id": "qkbdZzQA4DIu"
      },
      "execution_count": 50,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Логарифмирование"
      ],
      "metadata": {
        "id": "shbf8yM1z02P"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Хоть деревянным моделям и не важно распределение непрерывных величин, я решил перестраховаться и логарифмировать некоторые величины для преобразования в нормально распределённые.  "
      ],
      "metadata": {
        "id": "qqPkM7TSz40_"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "plt.subplot(131)\n",
        "sns.histplot(df['triglyceride'], color='blue')\n",
        "plt.title('Triglyceride')\n",
        "plt.xlabel('Values')\n",
        "plt.ylabel('Frequency')\n",
        "\n",
        "plt.subplot(132)\n",
        "sns.histplot(df['gtp'], color='green')\n",
        "plt.title('GTP')\n",
        "plt.xlabel('Values')\n",
        "plt.ylabel('Frequency')\n",
        "\n",
        "plt.subplot(133)\n",
        "sns.histplot(df['alt'], color='red')\n",
        "plt.title('ALT')\n",
        "plt.xlabel('Values')\n",
        "plt.ylabel('Frequency')\n",
        "\n",
        "plt.tight_layout()\n",
        "\n",
        "plt.show()\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 675
        },
        "id": "bsEO8JK30qEZ",
        "outputId": "7112633c-a510-4a7c-ce6d-afa40cba4270"
      },
      "execution_count": 51,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1600x900 with 3 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df['log_tr'] = df.triglyceride.apply(np.log)\n",
        "df['log_gtp'] = df.gtp.apply(np.log)\n",
        "df['log_log_alt'] = np.log(np.log(df.alt))"
      ],
      "metadata": {
        "id": "0RrEQHaa11dZ"
      },
      "execution_count": 52,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "plt.subplot(131)\n",
        "sns.histplot(df['log_tr'], color='blue')\n",
        "plt.title('Log_Triglyceride')\n",
        "plt.xlabel('Values')\n",
        "plt.ylabel('Frequency')\n",
        "\n",
        "plt.subplot(132)\n",
        "sns.histplot(df['log_gtp'], color='green')\n",
        "plt.title('Log_GTP')\n",
        "plt.xlabel('Values')\n",
        "plt.ylabel('Frequency')\n",
        "\n",
        "plt.subplot(133)\n",
        "sns.histplot(df['log_log_alt'], color='red')\n",
        "plt.title('Log_Log_ALT')\n",
        "plt.xlabel('Values')\n",
        "plt.ylabel('Frequency')\n",
        "\n",
        "plt.tight_layout()\n",
        "\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 675
        },
        "id": "trzRjCZW2Ql2",
        "outputId": "274068ed-9218-43f2-ff8c-5984b9857453"
      },
      "execution_count": 53,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1600x900 with 3 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "evaluate_features(['triglyceride', 'gtp', 'alt',\n",
        "                   'log_tr', 'log_gtp', 'log_log_alt'])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 254
        },
        "id": "A_pek0jK2g0W",
        "outputId": "34564450-6173-49f6-f7ca-b23ca422167d"
      },
      "execution_count": 54,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "interval columns not set, guessing: ['triglyceride', 'gtp', 'alt', 'log_tr', 'log_gtp', 'log_log_alt', 'smoking']\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "              smoking\n",
              "log_gtp          0.23\n",
              "log_tr           0.19\n",
              "triglyceride     0.19\n",
              "gtp              0.14\n",
              "log_log_alt      0.02\n",
              "alt              0.01"
            ],
            "text/html": [
              "\n",
              "\n",
              "  <div id=\"df-d3f83a39-24fe-4928-b3a5-0c32db65744a\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>smoking</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>log_gtp</th>\n",
              "      <td>0.23</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>log_tr</th>\n",
              "      <td>0.19</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>triglyceride</th>\n",
              "      <td>0.19</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>gtp</th>\n",
              "      <td>0.14</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>log_log_alt</th>\n",
              "      <td>0.02</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>alt</th>\n",
              "      <td>0.01</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-d3f83a39-24fe-4928-b3a5-0c32db65744a')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "\n",
              "\n",
              "\n",
              "    <div id=\"df-140326ac-4c5b-4f69-99c7-967fa05feb1f\">\n",
              "      <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-140326ac-4c5b-4f69-99c7-967fa05feb1f')\"\n",
              "              title=\"Suggest charts.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "      </button>\n",
              "    </div>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "    background-color: #E8F0FE;\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: #1967D2;\n",
              "    height: 32px;\n",
              "    padding: 0 0 0 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: #E2EBFA;\n",
              "    box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: #174EA6;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "    background-color: #3B4455;\n",
              "    fill: #D2E3FC;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart:hover {\n",
              "    background-color: #434B5C;\n",
              "    box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "    filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "    fill: #FFFFFF;\n",
              "  }\n",
              "</style>\n",
              "\n",
              "    <script>\n",
              "      async function quickchart(key) {\n",
              "        const containerElement = document.querySelector('#' + key);\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      }\n",
              "    </script>\n",
              "\n",
              "      <script>\n",
              "\n",
              "function displayQuickchartButton(domScope) {\n",
              "  let quickchartButtonEl =\n",
              "    domScope.querySelector('#df-140326ac-4c5b-4f69-99c7-967fa05feb1f button.colab-df-quickchart');\n",
              "  quickchartButtonEl.style.display =\n",
              "    google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "}\n",
              "\n",
              "        displayQuickchartButton(document);\n",
              "      </script>\n",
              "      <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-d3f83a39-24fe-4928-b3a5-0c32db65744a button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-d3f83a39-24fe-4928-b3a5-0c32db65744a');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 54
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "test['log_tr'] = test.triglyceride.apply(np.log)\n",
        "test['log_gtp'] = test.gtp.apply(np.log)\n",
        "test['log_log_alt'] = np.log(np.log(test.alt))"
      ],
      "metadata": {
        "id": "zsctpJsF27Xz"
      },
      "execution_count": 55,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## *Binning*"
      ],
      "metadata": {
        "id": "9nEOxQV1y_sX"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Для уменьшения шума в непрерывных величинах я поигрался с `pd.qcut` и выбрал такое количество квантилей, чтобы коэффициент *phik* был выше чем у исходного предиктора."
      ],
      "metadata": {
        "id": "0wYkWPUOzEMI"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "df['gtp_bin'] = pd.qcut(df['log_gtp'], q=4, labels=False)\n",
        "df['tr_bin'] = pd.qcut(df['log_tr'], q=4, labels=False)\n",
        "df['hem_bin'] = pd.qcut(df['hemoglobin'], q=6, labels=False)"
      ],
      "metadata": {
        "id": "zw1w2N_84paJ"
      },
      "execution_count": 56,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "evaluate_features(['log_gtp', 'log_tr', 'hemoglobin',\n",
        "                   'gtp_bin', 'tr_bin', 'hem_bin'])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 254
        },
        "id": "_6z4ngFr4864",
        "outputId": "4c24bfb3-8868-41d8-a481-f13f4366340e"
      },
      "execution_count": 57,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "interval columns not set, guessing: ['log_gtp', 'log_tr', 'hemoglobin', 'gtp_bin', 'tr_bin', 'hem_bin', 'smoking']\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "            smoking\n",
              "gtp_bin        0.26\n",
              "log_gtp        0.23\n",
              "tr_bin         0.22\n",
              "log_tr         0.19\n",
              "hem_bin        0.12\n",
              "hemoglobin     0.11"
            ],
            "text/html": [
              "\n",
              "\n",
              "  <div id=\"df-f5670ee4-b016-4684-b476-5e2df7424f52\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>smoking</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>gtp_bin</th>\n",
              "      <td>0.26</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>log_gtp</th>\n",
              "      <td>0.23</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>tr_bin</th>\n",
              "      <td>0.22</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>log_tr</th>\n",
              "      <td>0.19</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>hem_bin</th>\n",
              "      <td>0.12</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>hemoglobin</th>\n",
              "      <td>0.11</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-f5670ee4-b016-4684-b476-5e2df7424f52')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "\n",
              "\n",
              "\n",
              "    <div id=\"df-0733a241-4ccb-4037-84bf-619fe2195c00\">\n",
              "      <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-0733a241-4ccb-4037-84bf-619fe2195c00')\"\n",
              "              title=\"Suggest charts.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "      </button>\n",
              "    </div>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "    background-color: #E8F0FE;\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: #1967D2;\n",
              "    height: 32px;\n",
              "    padding: 0 0 0 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: #E2EBFA;\n",
              "    box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: #174EA6;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "    background-color: #3B4455;\n",
              "    fill: #D2E3FC;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart:hover {\n",
              "    background-color: #434B5C;\n",
              "    box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "    filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "    fill: #FFFFFF;\n",
              "  }\n",
              "</style>\n",
              "\n",
              "    <script>\n",
              "      async function quickchart(key) {\n",
              "        const containerElement = document.querySelector('#' + key);\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      }\n",
              "    </script>\n",
              "\n",
              "      <script>\n",
              "\n",
              "function displayQuickchartButton(domScope) {\n",
              "  let quickchartButtonEl =\n",
              "    domScope.querySelector('#df-0733a241-4ccb-4037-84bf-619fe2195c00 button.colab-df-quickchart');\n",
              "  quickchartButtonEl.style.display =\n",
              "    google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "}\n",
              "\n",
              "        displayQuickchartButton(document);\n",
              "      </script>\n",
              "      <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-f5670ee4-b016-4684-b476-5e2df7424f52 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-f5670ee4-b016-4684-b476-5e2df7424f52');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 57
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "test['gtp_bin'] = pd.qcut(test['log_gtp'], q=4, labels=False)\n",
        "test['tr_bin'] = pd.qcut(test['log_tr'], q=4, labels=False)\n",
        "test['hem_bin'] = pd.qcut(test['hemoglobin'], q=6, labels=False)"
      ],
      "metadata": {
        "id": "4mQRGwCX5H6t"
      },
      "execution_count": 58,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Я также пробовал передавать эти величины *CatBoost*-у как категориальные, но это не принесло значительной пользы."
      ],
      "metadata": {
        "id": "Dm5bPLCl5LFK"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Удаляем ненужное"
      ],
      "metadata": {
        "id": "QydIxxaBHF9g"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "df = df.drop([\n",
        "    'triglyceride', 'gtp', 'alt', 'eyesight_left', 'eyesight_right',\n",
        "    'hearing_left', 'hearing_right'\n",
        "],\n",
        "             axis=1)\n",
        "\n",
        "test = test.drop([\n",
        "    'triglyceride', 'gtp', 'alt', 'eyesight_left', 'eyesight_right',\n",
        "    'hearing_left', 'hearing_right'\n",
        "],\n",
        "             axis=1)\n"
      ],
      "metadata": {
        "id": "sbvc3ZfEFfj5"
      },
      "execution_count": 59,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Итог преобразования"
      ],
      "metadata": {
        "id": "2es7urQc5oeR"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "evaluate_features(df.drop(['id', 'smoking'], axis=1).columns.to_list())"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "F06hgW1aFo7w",
        "outputId": "9e5b340f-badd-4a83-9a6a-08bf26e3bce3"
      },
      "execution_count": 60,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "interval columns not set, guessing: ['age', 'height_cm', 'weight_kg', 'waist_cm', 'systolic', 'relaxation', 'fasting_blood_sugar', 'cholesterol', 'hdl', 'ldl', 'hemoglobin', 'urine_protein', 'serum_creatinine', 'ast', 'dental_caries', 'tartar', 'chol_hdl', 'ldl_hdl', 'map', 'wh_ratio', 'hearing_mean', 'eyesight_mean', 'bmi', 'ast_alt', 'log_tr', 'log_gtp', 'log_log_alt', 'gtp_bin', 'tr_bin', 'hem_bin', 'smoking']\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                     smoking\n",
              "gtp_bin                 0.26\n",
              "log_gtp                 0.23\n",
              "tr_bin                  0.22\n",
              "middle_aged             0.20\n",
              "log_tr                  0.19\n",
              "age                     0.18\n",
              "senior                  0.12\n",
              "hem_bin                 0.12\n",
              "hemoglobin              0.11\n",
              "tartar                  0.10\n",
              "dental_caries           0.09\n",
              "bmi                     0.07\n",
              "waist_cm                0.06\n",
              "normal_weight           0.06\n",
              "serum_creatinine        0.06\n",
              "hdl                     0.05\n",
              "dental_health           0.05\n",
              "bmi_cat                 0.05\n",
              "weight_kg               0.04\n",
              "fasting_blood_sugar     0.04\n",
              "height_cm               0.04\n",
              "wh_ratio                0.03\n",
              "chol_hdl_cat            0.03\n",
              "urine_protein           0.03\n",
              "ast_alt                 0.03\n",
              "map_cat                 0.02\n",
              "relaxation              0.02\n",
              "ldl                     0.02\n",
              "chol_hdl                0.02\n",
              "ldl_hdl                 0.02\n",
              "wh_cat                  0.02\n",
              "systolic                0.02\n",
              "cholesterol             0.02\n",
              "log_log_alt             0.02\n",
              "map                     0.02\n",
              "hearing_mean            0.01\n",
              "eyesight_mean           0.01\n",
              "ast                     0.00"
            ],
            "text/html": [
              "\n",
              "\n",
              "  <div id=\"df-4040d6f7-36c0-4b24-8692-b58291446202\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>smoking</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>gtp_bin</th>\n",
              "      <td>0.26</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>log_gtp</th>\n",
              "      <td>0.23</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>tr_bin</th>\n",
              "      <td>0.22</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>middle_aged</th>\n",
              "      <td>0.20</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>log_tr</th>\n",
              "      <td>0.19</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>age</th>\n",
              "      <td>0.18</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>senior</th>\n",
              "      <td>0.12</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>hem_bin</th>\n",
              "      <td>0.12</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>hemoglobin</th>\n",
              "      <td>0.11</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>tartar</th>\n",
              "      <td>0.10</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>dental_caries</th>\n",
              "      <td>0.09</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>bmi</th>\n",
              "      <td>0.07</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>waist_cm</th>\n",
              "      <td>0.06</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>normal_weight</th>\n",
              "      <td>0.06</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>serum_creatinine</th>\n",
              "      <td>0.06</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>hdl</th>\n",
              "      <td>0.05</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>dental_health</th>\n",
              "      <td>0.05</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>bmi_cat</th>\n",
              "      <td>0.05</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>weight_kg</th>\n",
              "      <td>0.04</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>fasting_blood_sugar</th>\n",
              "      <td>0.04</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>height_cm</th>\n",
              "      <td>0.04</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>wh_ratio</th>\n",
              "      <td>0.03</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>chol_hdl_cat</th>\n",
              "      <td>0.03</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>urine_protein</th>\n",
              "      <td>0.03</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>ast_alt</th>\n",
              "      <td>0.03</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>map_cat</th>\n",
              "      <td>0.02</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>relaxation</th>\n",
              "      <td>0.02</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>ldl</th>\n",
              "      <td>0.02</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>chol_hdl</th>\n",
              "      <td>0.02</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>ldl_hdl</th>\n",
              "      <td>0.02</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>wh_cat</th>\n",
              "      <td>0.02</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>systolic</th>\n",
              "      <td>0.02</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>cholesterol</th>\n",
              "      <td>0.02</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>log_log_alt</th>\n",
              "      <td>0.02</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>map</th>\n",
              "      <td>0.02</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>hearing_mean</th>\n",
              "      <td>0.01</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>eyesight_mean</th>\n",
              "      <td>0.01</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>ast</th>\n",
              "      <td>0.00</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-4040d6f7-36c0-4b24-8692-b58291446202')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "\n",
              "\n",
              "\n",
              "    <div id=\"df-90015ede-d7fe-435a-9087-c92f374c5464\">\n",
              "      <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-90015ede-d7fe-435a-9087-c92f374c5464')\"\n",
              "              title=\"Suggest charts.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "      </button>\n",
              "    </div>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "    background-color: #E8F0FE;\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: #1967D2;\n",
              "    height: 32px;\n",
              "    padding: 0 0 0 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: #E2EBFA;\n",
              "    box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: #174EA6;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "    background-color: #3B4455;\n",
              "    fill: #D2E3FC;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart:hover {\n",
              "    background-color: #434B5C;\n",
              "    box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "    filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "    fill: #FFFFFF;\n",
              "  }\n",
              "</style>\n",
              "\n",
              "    <script>\n",
              "      async function quickchart(key) {\n",
              "        const containerElement = document.querySelector('#' + key);\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      }\n",
              "    </script>\n",
              "\n",
              "      <script>\n",
              "\n",
              "function displayQuickchartButton(domScope) {\n",
              "  let quickchartButtonEl =\n",
              "    domScope.querySelector('#df-90015ede-d7fe-435a-9087-c92f374c5464 button.colab-df-quickchart');\n",
              "  quickchartButtonEl.style.display =\n",
              "    google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "}\n",
              "\n",
              "        displayQuickchartButton(document);\n",
              "      </script>\n",
              "      <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-4040d6f7-36c0-4b24-8692-b58291446202 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-4040d6f7-36c0-4b24-8692-b58291446202');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 60
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Предобработка для *ML*"
      ],
      "metadata": {
        "id": "4t-v0VmH5xnv"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "X = df.drop('id', axis='columns')\n",
        "y = X.pop('smoking')\n",
        "test_ids = test.pop('id')"
      ],
      "metadata": {
        "id": "Pspakb5NFrka"
      },
      "execution_count": 61,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "assert all(X.columns == test.columns)"
      ],
      "metadata": {
        "id": "yrB-PHOBqwY4"
      },
      "execution_count": 62,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "cat_cols = ['dental_caries', 'tartar', 'middle_aged', 'senior', 'normal_weight',\n",
        "            'bmi_cat', 'dental_health', 'map_cat', 'wh_cat', 'chol_hdl_cat']"
      ],
      "metadata": {
        "id": "uqtO3MCZqxHq"
      },
      "execution_count": 63,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "X[cat_cols] = X[cat_cols].astype('str')\n",
        "test[cat_cols] = test[cat_cols].astype('str')\n",
        "cat_cols"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Gpc-Mu6frD8f",
        "outputId": "3b7cd03a-f7f6-4520-85ac-67f4a9d68eda"
      },
      "execution_count": 64,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['dental_caries',\n",
              " 'tartar',\n",
              " 'middle_aged',\n",
              " 'senior',\n",
              " 'normal_weight',\n",
              " 'bmi_cat',\n",
              " 'dental_health',\n",
              " 'map_cat',\n",
              " 'wh_cat',\n",
              " 'chol_hdl_cat']"
            ]
          },
          "metadata": {},
          "execution_count": 64
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Предобработка для случайного леса"
      ],
      "metadata": {
        "id": "LSSGbBEKR8jy"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Масштабирование"
      ],
      "metadata": {
        "id": "n1zVopCSGIE-"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Для того чтобы генерировать *interaction features*, сначала стандартизирую их чтобы их масштаб не влиял на их важность."
      ],
      "metadata": {
        "id": "jgTnnAf_GM1l"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "st_scaler = StandardScaler()"
      ],
      "metadata": {
        "id": "xNykW7UUSJUg"
      },
      "execution_count": 65,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "continuous_cols = [col for col in X.columns if col not in cat_cols]\n",
        "continuous_cols"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "LIxUKgQiSPLi",
        "outputId": "1cf32aa3-5443-4b2a-adaf-5fbe6d1537ec"
      },
      "execution_count": 66,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['age',\n",
              " 'height_cm',\n",
              " 'weight_kg',\n",
              " 'waist_cm',\n",
              " 'systolic',\n",
              " 'relaxation',\n",
              " 'fasting_blood_sugar',\n",
              " 'cholesterol',\n",
              " 'hdl',\n",
              " 'ldl',\n",
              " 'hemoglobin',\n",
              " 'urine_protein',\n",
              " 'serum_creatinine',\n",
              " 'ast',\n",
              " 'chol_hdl',\n",
              " 'ldl_hdl',\n",
              " 'map',\n",
              " 'wh_ratio',\n",
              " 'hearing_mean',\n",
              " 'eyesight_mean',\n",
              " 'bmi',\n",
              " 'ast_alt',\n",
              " 'log_tr',\n",
              " 'log_gtp',\n",
              " 'log_log_alt',\n",
              " 'gtp_bin',\n",
              " 'tr_bin',\n",
              " 'hem_bin']"
            ]
          },
          "metadata": {},
          "execution_count": 66
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "assert len(continuous_cols) + len(cat_cols) == X.shape[1]"
      ],
      "metadata": {
        "id": "3ujX9FALSmSK"
      },
      "execution_count": 67,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "X_rfc = X.copy()\n",
        "test_rfc = test.copy()"
      ],
      "metadata": {
        "id": "xEZjzC4E2_iB"
      },
      "execution_count": 68,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "st_scaler.fit(X_rfc[continuous_cols])\n",
        "pass"
      ],
      "metadata": {
        "id": "EnhZ8EGdTFEX"
      },
      "execution_count": 69,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def standard_scale(X: pd.DataFrame) -> pd.DataFrame:\n",
        "    '''\n",
        "    Standard Scaling с возвратом данных в формате датафрейма,\n",
        "    со смысловыми столбцами\n",
        "    '''\n",
        "    return pd.DataFrame(st_scaler.transform(X),\n",
        "                        columns=X.columns,\n",
        "                        index=X.index)"
      ],
      "metadata": {
        "id": "lhHk3FeuSBpI"
      },
      "execution_count": 70,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "X_rfc[continuous_cols] = standard_scale(X_rfc[continuous_cols])\n",
        "test_rfc[continuous_cols] = standard_scale(test_rfc[continuous_cols])"
      ],
      "metadata": {
        "id": "c2dMjSeJS3jy"
      },
      "execution_count": 71,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### *Interaction Features*"
      ],
      "metadata": {
        "id": "8Ooj3gujGubj"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Создаю все возможные произведения двух предикторов с помощью `PolynomialFeatures`:"
      ],
      "metadata": {
        "id": "c-1TK8rYGzaf"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "polynom = PolynomialFeatures(interaction_only=True, include_bias=False)\n",
        "\n",
        "X_poly = \\\n",
        "pd.DataFrame(polynom.fit_transform(X_rfc[continuous_cols]))\n",
        "X_rfc = X_poly.join(X[cat_cols])\n",
        "\n",
        "test_poly = \\\n",
        "pd.DataFrame(polynom.fit_transform(test_rfc[continuous_cols]))\n",
        "test_rfc = test_poly.join(test_rfc[cat_cols])"
      ],
      "metadata": {
        "id": "qzbi3o3DS4A7"
      },
      "execution_count": 72,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Заново присоединяю целевой признак для анализа корреляции:"
      ],
      "metadata": {
        "id": "R0M_UwrrHGv3"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "X_exper = X_rfc.join(y)\n",
        "X_exper.columns[-1]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "id": "rhTVZFn0S4MM",
        "outputId": "5f4499fc-44ea-40f0-c9b3-9f2e1e3990db"
      },
      "execution_count": 73,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "'smoking'"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            }
          },
          "metadata": {},
          "execution_count": 73
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Выбор признаков"
      ],
      "metadata": {
        "id": "OLX3ThaMIJAx"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Не нашёл как считать матрицу *phik* только для одного столбца, поэтому пришлось считать для всей матрицы $417$ * $417$  ¯\\\\_(ツ)_/¯."
      ],
      "metadata": {
        "id": "3KW5CzntHN4w"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "table = X_exper\\\n",
        "    .phik_matrix()['smoking'].to_frame()\\\n",
        "    .sort_values('smoking', ascending=False).round(2)\n",
        "display(table.iloc[1:50, :])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "h_dEo8kxWiup",
        "outputId": "c9be1535-1937-4e3b-8e76-ae7cfe3867be"
      },
      "execution_count": 74,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "interval columns not set, guessing: [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 128, 129, 130, 131, 132, 133, 134, 135, 136, 137, 138, 139, 140, 141, 142, 143, 144, 145, 146, 147, 148, 149, 150, 151, 152, 153, 154, 155, 156, 157, 158, 159, 160, 161, 162, 163, 164, 165, 166, 167, 168, 169, 170, 171, 172, 173, 174, 175, 176, 177, 178, 179, 180, 181, 182, 183, 184, 185, 186, 187, 188, 189, 190, 191, 192, 193, 194, 195, 196, 197, 198, 199, 200, 201, 202, 203, 204, 205, 206, 207, 208, 209, 210, 211, 212, 213, 214, 215, 216, 217, 218, 219, 220, 221, 222, 223, 224, 225, 226, 227, 228, 229, 230, 231, 232, 233, 234, 235, 236, 237, 238, 239, 240, 241, 242, 243, 244, 245, 246, 247, 248, 249, 250, 251, 252, 253, 254, 255, 256, 257, 258, 259, 260, 261, 262, 263, 264, 265, 266, 267, 268, 269, 270, 271, 272, 273, 274, 275, 276, 277, 278, 279, 280, 281, 282, 283, 284, 285, 286, 287, 288, 289, 290, 291, 292, 293, 294, 295, 296, 297, 298, 299, 300, 301, 302, 303, 304, 305, 306, 307, 308, 309, 310, 311, 312, 313, 314, 315, 316, 317, 318, 319, 320, 321, 322, 323, 324, 325, 326, 327, 328, 329, 330, 331, 332, 333, 334, 335, 336, 337, 338, 339, 340, 341, 342, 343, 344, 345, 346, 347, 348, 349, 350, 351, 352, 353, 354, 355, 356, 357, 358, 359, 360, 361, 362, 363, 364, 365, 366, 367, 368, 369, 370, 371, 372, 373, 374, 375, 376, 377, 378, 379, 380, 381, 382, 383, 384, 385, 386, 387, 388, 389, 390, 391, 392, 393, 394, 395, 396, 397, 398, 399, 400, 401, 402, 403, 404, 405, 'smoking']\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "               smoking\n",
              "25                0.26\n",
              "23                0.23\n",
              "26                0.22\n",
              "middle_aged       0.20\n",
              "22                0.19\n",
              "0                 0.18\n",
              "368               0.17\n",
              "367               0.15\n",
              "397               0.13\n",
              "404               0.13\n",
              "398               0.13\n",
              "senior            0.12\n",
              "27                0.12\n",
              "399               0.12\n",
              "53                0.11\n",
              "10                0.11\n",
              "405               0.11\n",
              "391               0.11\n",
              "52                0.11\n",
              "335               0.11\n",
              "248               0.11\n",
              "382               0.11\n",
              "28                0.11\n",
              "tartar            0.10\n",
              "125               0.10\n",
              "101               0.10\n",
              "380               0.10\n",
              "54                0.09\n",
              "dental_caries     0.09\n",
              "31                0.09\n",
              "323               0.09\n",
              "356               0.09\n",
              "32                0.09\n",
              "43                0.09\n",
              "38                0.08\n",
              "358               0.08\n",
              "127               0.08\n",
              "30                0.08\n",
              "393               0.08\n",
              "41                0.08\n",
              "384               0.07\n",
              "211               0.07\n",
              "29                0.07\n",
              "265               0.07\n",
              "259               0.07\n",
              "50                0.07\n",
              "327               0.07\n",
              "129               0.07\n",
              "20                0.07"
            ],
            "text/html": [
              "\n",
              "\n",
              "  <div id=\"df-360a1dad-b376-462d-983a-28c52ac6a6ce\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>smoking</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>25</th>\n",
              "      <td>0.26</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>23</th>\n",
              "      <td>0.23</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>26</th>\n",
              "      <td>0.22</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>middle_aged</th>\n",
              "      <td>0.20</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>22</th>\n",
              "      <td>0.19</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0.18</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>368</th>\n",
              "      <td>0.17</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>367</th>\n",
              "      <td>0.15</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>397</th>\n",
              "      <td>0.13</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>404</th>\n",
              "      <td>0.13</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>398</th>\n",
              "      <td>0.13</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>senior</th>\n",
              "      <td>0.12</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>27</th>\n",
              "      <td>0.12</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>399</th>\n",
              "      <td>0.12</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>53</th>\n",
              "      <td>0.11</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10</th>\n",
              "      <td>0.11</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>405</th>\n",
              "      <td>0.11</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>391</th>\n",
              "      <td>0.11</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>52</th>\n",
              "      <td>0.11</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>335</th>\n",
              "      <td>0.11</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>248</th>\n",
              "      <td>0.11</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>382</th>\n",
              "      <td>0.11</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>28</th>\n",
              "      <td>0.11</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>tartar</th>\n",
              "      <td>0.10</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>125</th>\n",
              "      <td>0.10</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>101</th>\n",
              "      <td>0.10</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>380</th>\n",
              "      <td>0.10</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>54</th>\n",
              "      <td>0.09</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>dental_caries</th>\n",
              "      <td>0.09</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>31</th>\n",
              "      <td>0.09</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>323</th>\n",
              "      <td>0.09</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>356</th>\n",
              "      <td>0.09</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>32</th>\n",
              "      <td>0.09</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>43</th>\n",
              "      <td>0.09</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>38</th>\n",
              "      <td>0.08</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>358</th>\n",
              "      <td>0.08</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>127</th>\n",
              "      <td>0.08</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>30</th>\n",
              "      <td>0.08</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>393</th>\n",
              "      <td>0.08</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>41</th>\n",
              "      <td>0.08</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>384</th>\n",
              "      <td>0.07</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>211</th>\n",
              "      <td>0.07</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>29</th>\n",
              "      <td>0.07</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>265</th>\n",
              "      <td>0.07</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>259</th>\n",
              "      <td>0.07</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>50</th>\n",
              "      <td>0.07</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>327</th>\n",
              "      <td>0.07</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>129</th>\n",
              "      <td>0.07</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>20</th>\n",
              "      <td>0.07</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-360a1dad-b376-462d-983a-28c52ac6a6ce')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "\n",
              "\n",
              "\n",
              "    <div id=\"df-f2eedeac-d18c-47a2-b65c-3584be83f2cb\">\n",
              "      <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-f2eedeac-d18c-47a2-b65c-3584be83f2cb')\"\n",
              "              title=\"Suggest charts.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "      </button>\n",
              "    </div>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "    background-color: #E8F0FE;\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: #1967D2;\n",
              "    height: 32px;\n",
              "    padding: 0 0 0 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: #E2EBFA;\n",
              "    box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: #174EA6;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "    background-color: #3B4455;\n",
              "    fill: #D2E3FC;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart:hover {\n",
              "    background-color: #434B5C;\n",
              "    box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "    filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "    fill: #FFFFFF;\n",
              "  }\n",
              "</style>\n",
              "\n",
              "    <script>\n",
              "      async function quickchart(key) {\n",
              "        const containerElement = document.querySelector('#' + key);\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      }\n",
              "    </script>\n",
              "\n",
              "      <script>\n",
              "\n",
              "function displayQuickchartButton(domScope) {\n",
              "  let quickchartButtonEl =\n",
              "    domScope.querySelector('#df-f2eedeac-d18c-47a2-b65c-3584be83f2cb button.colab-df-quickchart');\n",
              "  quickchartButtonEl.style.display =\n",
              "    google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "}\n",
              "\n",
              "        displayQuickchartButton(document);\n",
              "      </script>\n",
              "      <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-360a1dad-b376-462d-983a-28c52ac6a6ce button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-360a1dad-b376-462d-983a-28c52ac6a6ce');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Создаю список из $100$ наиболее коллерирующих с целевым признаком фичей.  \n",
        "Такой выбор с гиперпараметром случайного леса `max_features='squared'` (по умолчанию) работает лучше, чем выбор меньшего кол-ва признаков с гиперпараметром `max_features=None`."
      ],
      "metadata": {
        "id": "3079UT3wI2Ui"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "strong_cols = table.iloc[1:101, :].index.to_list()"
      ],
      "metadata": {
        "id": "64voXi6afWeN"
      },
      "execution_count": 75,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "X_rfc = X_rfc[[col for col in strong_cols\n",
        "              if col not in cat_cols] + cat_cols]\n",
        "test_rfc = test_rfc[[col for col in strong_cols\n",
        "                    if col not in cat_cols] + cat_cols]"
      ],
      "metadata": {
        "id": "fTQIvSxQfl8B"
      },
      "execution_count": 76,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### *OHE*-Кодирование категориальных признаков"
      ],
      "metadata": {
        "id": "GgKXTafMJoN9"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "ohe_enc = OneHotEncoder(drop='first', sparse=False)\n",
        "ohe_enc.fit(X_rfc[cat_cols])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 129
        },
        "id": "ndjAk-RErtD9",
        "outputId": "48ceaebe-ef8d-4975-ef64-c304c6435dde"
      },
      "execution_count": 77,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/sklearn/preprocessing/_encoders.py:868: FutureWarning: `sparse` was renamed to `sparse_output` in version 1.2 and will be removed in 1.4. `sparse_output` is ignored unless you leave `sparse` to its default value.\n",
            "  warnings.warn(\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "OneHotEncoder(drop='first', sparse=False, sparse_output=False)"
            ],
            "text/html": [
              "<style>#sk-container-id-1 {color: black;background-color: white;}#sk-container-id-1 pre{padding: 0;}#sk-container-id-1 div.sk-toggleable {background-color: white;}#sk-container-id-1 label.sk-toggleable__label {cursor: pointer;display: block;width: 100%;margin-bottom: 0;padding: 0.3em;box-sizing: border-box;text-align: center;}#sk-container-id-1 label.sk-toggleable__label-arrow:before {content: \"▸\";float: left;margin-right: 0.25em;color: #696969;}#sk-container-id-1 label.sk-toggleable__label-arrow:hover:before {color: black;}#sk-container-id-1 div.sk-estimator:hover label.sk-toggleable__label-arrow:before {color: black;}#sk-container-id-1 div.sk-toggleable__content {max-height: 0;max-width: 0;overflow: hidden;text-align: left;background-color: #f0f8ff;}#sk-container-id-1 div.sk-toggleable__content pre {margin: 0.2em;color: black;border-radius: 0.25em;background-color: #f0f8ff;}#sk-container-id-1 input.sk-toggleable__control:checked~div.sk-toggleable__content {max-height: 200px;max-width: 100%;overflow: auto;}#sk-container-id-1 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {content: \"▾\";}#sk-container-id-1 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-1 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-1 input.sk-hidden--visually {border: 0;clip: rect(1px 1px 1px 1px);clip: rect(1px, 1px, 1px, 1px);height: 1px;margin: -1px;overflow: hidden;padding: 0;position: absolute;width: 1px;}#sk-container-id-1 div.sk-estimator {font-family: monospace;background-color: #f0f8ff;border: 1px dotted black;border-radius: 0.25em;box-sizing: border-box;margin-bottom: 0.5em;}#sk-container-id-1 div.sk-estimator:hover {background-color: #d4ebff;}#sk-container-id-1 div.sk-parallel-item::after {content: \"\";width: 100%;border-bottom: 1px solid gray;flex-grow: 1;}#sk-container-id-1 div.sk-label:hover label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-1 div.sk-serial::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: 0;}#sk-container-id-1 div.sk-serial {display: flex;flex-direction: column;align-items: center;background-color: white;padding-right: 0.2em;padding-left: 0.2em;position: relative;}#sk-container-id-1 div.sk-item {position: relative;z-index: 1;}#sk-container-id-1 div.sk-parallel {display: flex;align-items: stretch;justify-content: center;background-color: white;position: relative;}#sk-container-id-1 div.sk-item::before, #sk-container-id-1 div.sk-parallel-item::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: -1;}#sk-container-id-1 div.sk-parallel-item {display: flex;flex-direction: column;z-index: 1;position: relative;background-color: white;}#sk-container-id-1 div.sk-parallel-item:first-child::after {align-self: flex-end;width: 50%;}#sk-container-id-1 div.sk-parallel-item:last-child::after {align-self: flex-start;width: 50%;}#sk-container-id-1 div.sk-parallel-item:only-child::after {width: 0;}#sk-container-id-1 div.sk-dashed-wrapped {border: 1px dashed gray;margin: 0 0.4em 0.5em 0.4em;box-sizing: border-box;padding-bottom: 0.4em;background-color: white;}#sk-container-id-1 div.sk-label label {font-family: monospace;font-weight: bold;display: inline-block;line-height: 1.2em;}#sk-container-id-1 div.sk-label-container {text-align: center;}#sk-container-id-1 div.sk-container {/* jupyter's `normalize.less` sets `[hidden] { display: none; }` but bootstrap.min.css set `[hidden] { display: none !important; }` so we also need the `!important` here to be able to override the default hidden behavior on the sphinx rendered scikit-learn.org. See: https://github.com/scikit-learn/scikit-learn/issues/21755 */display: inline-block !important;position: relative;}#sk-container-id-1 div.sk-text-repr-fallback {display: none;}</style><div id=\"sk-container-id-1\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>OneHotEncoder(drop=&#x27;first&#x27;, sparse=False, sparse_output=False)</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-1\" type=\"checkbox\" checked><label for=\"sk-estimator-id-1\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">OneHotEncoder</label><div class=\"sk-toggleable__content\"><pre>OneHotEncoder(drop=&#x27;first&#x27;, sparse=False, sparse_output=False)</pre></div></div></div></div></div>"
            ]
          },
          "metadata": {},
          "execution_count": 77
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def ohe_transform(df: pd.DataFrame, ohe_cols: list) -> pd.DataFrame:\n",
        "    '''\n",
        "    One-hot encoding с возвратом данных в формате датафрейма,\n",
        "    со смысловыми столбцами\n",
        "    '''\n",
        "    X = df.copy()\n",
        "    ohe_cats = pd.DataFrame(ohe_enc.transform(X[ohe_cols]))\n",
        "    # берём индекс для дальнейшего склеивания\n",
        "    ohe_cats.index = X[ohe_cols].index\n",
        "    # переименовываем колонки в соответствие со смыслом\n",
        "    ohe_cats.columns = ohe_enc.get_feature_names_out(ohe_cols)\n",
        "    # удаляем непреобразованные категориальные столбцы\n",
        "    X = X.drop(ohe_cols, axis=1)\n",
        "    X = X.join(ohe_cats)\n",
        "    return X"
      ],
      "metadata": {
        "id": "gOQMywUjrqGp"
      },
      "execution_count": 78,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "X_rfc = ohe_transform(X_rfc, cat_cols)\n",
        "test_rfc = ohe_transform(test_rfc, cat_cols)"
      ],
      "metadata": {
        "id": "gLiR1m_Lr5kj"
      },
      "execution_count": 79,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "X_rfc.columns = X_rfc.columns.astype(str)\n",
        "test_rfc.columns = test_rfc.columns.astype(str)\n",
        "print(X_rfc.shape, test_rfc.shape)\n",
        "assert all(X_rfc.columns == test_rfc.columns)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ay1HGI4Mr5aw",
        "outputId": "11f7fd17-3d4d-44fc-9a9e-396538fea466"
      },
      "execution_count": 80,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(13863, 111) (5942, 111)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Оптимизация *CatBoost*-а"
      ],
      "metadata": {
        "id": "atJabIfEF2Gi"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Напоминаю, что *CatBoost*-у даю только интерпретируемые фичи:"
      ],
      "metadata": {
        "id": "iznezgbxNhUo"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "X.columns"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "HPsC9aeTNpb8",
        "outputId": "388a46ca-b066-43c4-8f02-15fd77890ef8"
      },
      "execution_count": 81,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Index(['age', 'height_cm', 'weight_kg', 'waist_cm', 'systolic', 'relaxation',\n",
              "       'fasting_blood_sugar', 'cholesterol', 'hdl', 'ldl', 'hemoglobin',\n",
              "       'urine_protein', 'serum_creatinine', 'ast', 'dental_caries', 'tartar',\n",
              "       'chol_hdl', 'ldl_hdl', 'chol_hdl_cat', 'map', 'map_cat',\n",
              "       'dental_health', 'wh_ratio', 'wh_cat', 'hearing_mean', 'eyesight_mean',\n",
              "       'middle_aged', 'senior', 'normal_weight', 'bmi', 'bmi_cat', 'ast_alt',\n",
              "       'log_tr', 'log_gtp', 'log_log_alt', 'gtp_bin', 'tr_bin', 'hem_bin'],\n",
              "      dtype='object')"
            ]
          },
          "metadata": {},
          "execution_count": 81
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Подбор гиперпараметров"
      ],
      "metadata": {
        "id": "0ZyMa3gHPoNb"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Подбирал модель *CatBoost* с помощью байесовской оптимизации поиска оптимальных гиперпараметров в библиотеке *Optuna*.\n",
        "\n",
        "В процессе многих запусков заметил, что качество модели лучшее на максимальной глубине одного дерева около 9-10."
      ],
      "metadata": {
        "id": "NWfArTmiNsOz"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "data_pool = Pool(data=X, label=y, cat_features=cat_cols)"
      ],
      "metadata": {
        "id": "fSRQEaPfGWe4"
      },
      "execution_count": 82,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Целевая функция — среднее значение *ROC-AUC* на стратифицированной кросс-валидации с пятью фолдами."
      ],
      "metadata": {
        "id": "CyH64b5ROcf3"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def objective_cb(n_trials):\n",
        "    cb_params = {\n",
        "        'n_estimators': n_trials.suggest_int('n_estimators',\n",
        "                                             800, 1100, step=50),\n",
        "        'learning_rate': n_trials.suggest_float('learning_rate',\n",
        "                                                1e-3, 0.4, log=True),\n",
        "        'l2_leaf_reg': n_trials.suggest_float('l2_leaf_reg', 1, 10),\n",
        "        'max_depth': n_trials.suggest_int('max_depth', 9, 10),\n",
        "        'loss_function': 'Logloss',\n",
        "        'eval_metric': 'AUC',\n",
        "        'cat_features': cat_cols,\n",
        "        'verbose': 500,\n",
        "        'random_state': 42\n",
        "    }\n",
        "\n",
        "    cv_data = cv(\n",
        "    params = cb_params,\n",
        "    pool = data_pool,\n",
        "    fold_count=5,\n",
        "    shuffle=True,\n",
        "    partition_random_seed=42,\n",
        "    stratified=True,\n",
        "    verbose=False,\n",
        "    early_stopping_rounds=200\n",
        "    )\n",
        "    roc_auc = cv_data['test-AUC-mean'].max()\n",
        "\n",
        "    return roc_auc"
      ],
      "metadata": {
        "id": "l8yVwrdMGGAL"
      },
      "execution_count": 83,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Запускаю подбор на $20$ итерациях:"
      ],
      "metadata": {
        "id": "7t-JtrUjP9Dh"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "study_cb = optuna.create_study(\n",
        "    direction='maximize',\n",
        "    pruner=HyperbandPruner(),\n",
        "    sampler=optuna.samplers.TPESampler(seed=42))\n",
        "study_cb.optimize(objective_cb, n_trials=20)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "u-K0BrIlGXhq",
        "outputId": "328dd96e-5704-4223-c9ea-afa8ac43658c"
      },
      "execution_count": 84,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "[I 2023-07-29 11:56:26,356] A new study created in memory with name: no-name-fe189320-7408-4e9d-ac6e-d9c583537357\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Training on fold [0/5]\n",
            "\n",
            "bestTest = 0.6962889742\n",
            "bestIteration = 17\n",
            "\n",
            "Training on fold [1/5]\n",
            "\n",
            "bestTest = 0.7126452456\n",
            "bestIteration = 898\n",
            "\n",
            "Training on fold [2/5]\n",
            "\n",
            "bestTest = 0.7106166484\n",
            "bestIteration = 899\n",
            "\n",
            "Training on fold [3/5]\n",
            "\n",
            "bestTest = 0.7146280031\n",
            "bestIteration = 896\n",
            "\n",
            "Training on fold [4/5]\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "[I 2023-07-29 12:04:35,026] Trial 0 finished with value: 0.7101705374660174 and parameters: {'n_estimators': 900, 'learning_rate': 0.2977252440242358, 'l2_leaf_reg': 7.587945476302646, 'max_depth': 10}. Best is trial 0 with value: 0.7101705374660174.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "bestTest = 0.7189631232\n",
            "bestIteration = 870\n",
            "\n",
            "Training on fold [0/5]\n",
            "\n",
            "bestTest = 0.7085743335\n",
            "bestIteration = 849\n",
            "\n",
            "Training on fold [1/5]\n",
            "\n",
            "bestTest = 0.6920623265\n",
            "bestIteration = 847\n",
            "\n",
            "Training on fold [2/5]\n",
            "\n",
            "bestTest = 0.7075334065\n",
            "bestIteration = 846\n",
            "\n",
            "Training on fold [3/5]\n",
            "\n",
            "bestTest = 0.6964164622\n",
            "bestIteration = 844\n",
            "\n",
            "Training on fold [4/5]\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "[I 2023-07-29 12:12:19,427] Trial 1 finished with value: 0.702785532560877 and parameters: {'n_estimators': 850, 'learning_rate': 0.002546285519362238, 'l2_leaf_reg': 1.5227525095137953, 'max_depth': 10}. Best is trial 0 with value: 0.7101705374660174.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "bestTest = 0.7094363537\n",
            "bestIteration = 849\n",
            "\n",
            "Training on fold [0/5]\n",
            "\n",
            "bestTest = 0.7384529727\n",
            "bestIteration = 785\n",
            "\n",
            "Training on fold [1/5]\n",
            "\n",
            "bestTest = 0.7276111936\n",
            "bestIteration = 946\n",
            "\n",
            "Training on fold [2/5]\n",
            "\n",
            "bestTest = 0.7177958169\n",
            "bestIteration = 245\n",
            "\n",
            "Training on fold [3/5]\n",
            "\n",
            "bestTest = 0.7372691165\n",
            "bestIteration = 989\n",
            "\n",
            "Training on fold [4/5]\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "[I 2023-07-29 12:22:20,902] Trial 2 finished with value: 0.7329486329371905 and parameters: {'n_estimators': 1000, 'learning_rate': 0.06957405651599215, 'l2_leaf_reg': 1.185260448662222, 'max_depth': 10}. Best is trial 2 with value: 0.7329486329371905.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "bestTest = 0.7454856626\n",
            "bestIteration = 996\n",
            "\n",
            "Training on fold [0/5]\n",
            "\n",
            "bestTest = 0.7140242722\n",
            "bestIteration = 1049\n",
            "\n",
            "Training on fold [1/5]\n",
            "\n",
            "bestTest = 0.6992447227\n",
            "bestIteration = 1046\n",
            "\n",
            "Training on fold [2/5]\n",
            "\n",
            "bestTest = 0.7154468724\n",
            "bestIteration = 1048\n",
            "\n",
            "Training on fold [3/5]\n",
            "\n",
            "bestTest = 0.7075360049\n",
            "bestIteration = 1049\n",
            "\n",
            "Training on fold [4/5]\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "[I 2023-07-29 12:30:45,036] Trial 3 finished with value: 0.7105319890286876 and parameters: {'n_estimators': 1050, 'learning_rate': 0.0035687744787509814, 'l2_leaf_reg': 2.636424704863906, 'max_depth': 9}. Best is trial 2 with value: 0.7329486329371905.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "bestTest = 0.716486373\n",
            "bestIteration = 1045\n",
            "\n",
            "Training on fold [0/5]\n",
            "\n",
            "bestTest = 0.7323808986\n",
            "bestIteration = 808\n",
            "\n",
            "Training on fold [1/5]\n",
            "\n",
            "bestTest = 0.7215883416\n",
            "bestIteration = 898\n",
            "\n",
            "Training on fold [2/5]\n",
            "\n",
            "bestTest = 0.7282107353\n",
            "bestIteration = 771\n",
            "\n",
            "Training on fold [3/5]\n",
            "\n",
            "bestTest = 0.7276301343\n",
            "bestIteration = 860\n",
            "\n",
            "Training on fold [4/5]\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "[I 2023-07-29 12:38:29,043] Trial 4 finished with value: 0.7278331443515649 and parameters: {'n_estimators': 900, 'learning_rate': 0.0231978489335329, 'l2_leaf_reg': 4.887505167779041, 'max_depth': 9}. Best is trial 2 with value: 0.7329486329371905.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "bestTest = 0.7321654611\n",
            "bestIteration = 897\n",
            "\n",
            "Training on fold [0/5]\n",
            "\n",
            "bestTest = 0.70022271\n",
            "bestIteration = 998\n",
            "\n",
            "Training on fold [1/5]\n",
            "\n",
            "bestTest = 0.688945194\n",
            "bestIteration = 999\n",
            "\n",
            "Training on fold [2/5]\n",
            "\n",
            "bestTest = 0.7029266994\n",
            "bestIteration = 999\n",
            "\n",
            "Training on fold [3/5]\n",
            "\n",
            "bestTest = 0.6943950207\n",
            "bestIteration = 999\n",
            "\n",
            "Training on fold [4/5]\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "[I 2023-07-29 12:45:51,329] Trial 5 finished with value: 0.6983133284501774 and parameters: {'n_estimators': 1000, 'learning_rate': 0.002306595252621264, 'l2_leaf_reg': 3.629301836816963, 'max_depth': 9}. Best is trial 2 with value: 0.7329486329371905.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "bestTest = 0.7051432123\n",
            "bestIteration = 996\n",
            "\n",
            "Training on fold [0/5]\n",
            "\n",
            "bestTest = 0.7344441934\n",
            "bestIteration = 749\n",
            "\n",
            "Training on fold [1/5]\n",
            "\n",
            "bestTest = 0.7208040152\n",
            "bestIteration = 771\n",
            "\n",
            "Training on fold [2/5]\n",
            "\n",
            "bestTest = 0.7363356142\n",
            "bestIteration = 577\n",
            "\n",
            "Training on fold [3/5]\n",
            "\n",
            "bestTest = 0.7253987988\n",
            "bestIteration = 254\n",
            "\n",
            "Training on fold [4/5]\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "[I 2023-07-29 12:54:07,418] Trial 6 finished with value: 0.7293321157691399 and parameters: {'n_estimators': 950, 'learning_rate': 0.11042691903147239, 'l2_leaf_reg': 2.7970640394252375, 'max_depth': 10}. Best is trial 2 with value: 0.7329486329371905.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "bestTest = 0.7359298308\n",
            "bestIteration = 352\n",
            "\n",
            "Training on fold [0/5]\n",
            "\n",
            "bestTest = 0.6877517591\n",
            "bestIteration = 999\n",
            "\n",
            "Training on fold [1/5]\n",
            "\n",
            "bestTest = 0.6737694468\n",
            "bestIteration = 999\n",
            "\n",
            "Training on fold [2/5]\n",
            "\n",
            "bestTest = 0.6894043315\n",
            "bestIteration = 999\n",
            "\n",
            "Training on fold [3/5]\n",
            "\n",
            "bestTest = 0.6805751098\n",
            "bestIteration = 998\n",
            "\n",
            "Training on fold [4/5]\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "[I 2023-07-29 13:00:53,213] Trial 7 finished with value: 0.6848610674657618 and parameters: {'n_estimators': 1000, 'learning_rate': 0.001320890329148899, 'l2_leaf_reg': 6.467903667112945, 'max_depth': 9}. Best is trial 2 with value: 0.7329486329371905.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "bestTest = 0.6928506846\n",
            "bestIteration = 998\n",
            "\n",
            "Training on fold [0/5]\n",
            "\n",
            "bestTest = 0.7184865406\n",
            "bestIteration = 736\n",
            "\n",
            "Training on fold [1/5]\n",
            "\n",
            "bestTest = 0.7188649861\n",
            "bestIteration = 316\n",
            "\n",
            "Training on fold [2/5]\n",
            "\n",
            "bestTest = 0.7286496998\n",
            "bestIteration = 750\n",
            "\n",
            "Training on fold [3/5]\n",
            "\n",
            "bestTest = 0.7137343387\n",
            "bestIteration = 795\n",
            "\n",
            "Training on fold [4/5]\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "[I 2023-07-29 13:08:23,771] Trial 8 finished with value: 0.7209438136394428 and parameters: {'n_estimators': 800, 'learning_rate': 0.2944808736192367, 'l2_leaf_reg': 9.690688297671034, 'max_depth': 10}. Best is trial 2 with value: 0.7329486329371905.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "bestTest = 0.726974619\n",
            "bestIteration = 352\n",
            "\n",
            "Training on fold [0/5]\n",
            "\n",
            "bestTest = 0.6893664063\n",
            "bestIteration = 899\n",
            "\n",
            "Training on fold [1/5]\n",
            "\n",
            "bestTest = 0.678465722\n",
            "bestIteration = 899\n",
            "\n",
            "Training on fold [2/5]\n",
            "\n",
            "bestTest = 0.6931064812\n",
            "bestIteration = 898\n",
            "\n",
            "Training on fold [3/5]\n",
            "\n",
            "bestTest = 0.6848698657\n",
            "bestIteration = 898\n",
            "\n",
            "Training on fold [4/5]\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "[I 2023-07-29 13:14:35,881] Trial 9 finished with value: 0.6887435225324208 and parameters: {'n_estimators': 900, 'learning_rate': 0.0017953482393100694, 'l2_leaf_reg': 7.158097238609412, 'max_depth': 9}. Best is trial 2 with value: 0.7329486329371905.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "bestTest = 0.6980019698\n",
            "bestIteration = 895\n",
            "\n",
            "Training on fold [0/5]\n",
            "\n",
            "bestTest = 0.7314819895\n",
            "bestIteration = 310\n",
            "\n",
            "Training on fold [1/5]\n",
            "\n",
            "bestTest = 0.7248999419\n",
            "bestIteration = 514\n",
            "\n",
            "Training on fold [2/5]\n",
            "\n",
            "bestTest = 0.7403589181\n",
            "bestIteration = 980\n",
            "\n",
            "Training on fold [3/5]\n",
            "\n",
            "bestTest = 0.7279627357\n",
            "bestIteration = 471\n",
            "\n",
            "Training on fold [4/5]\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "[I 2023-07-29 13:23:48,961] Trial 10 finished with value: 0.7333003166074498 and parameters: {'n_estimators': 1100, 'learning_rate': 0.043760845570680794, 'l2_leaf_reg': 1.0679258738466366, 'max_depth': 10}. Best is trial 10 with value: 0.7333003166074498.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "bestTest = 0.744506426\n",
            "bestIteration = 1088\n",
            "\n",
            "Training on fold [0/5]\n",
            "\n",
            "bestTest = 0.7284326383\n",
            "bestIteration = 289\n",
            "\n",
            "Training on fold [1/5]\n",
            "\n",
            "bestTest = 0.7262273255\n",
            "bestIteration = 920\n",
            "\n",
            "Training on fold [2/5]\n",
            "\n",
            "bestTest = 0.7301812343\n",
            "bestIteration = 1098\n",
            "\n",
            "Training on fold [3/5]\n",
            "\n",
            "bestTest = 0.7222713769\n",
            "bestIteration = 324\n",
            "\n",
            "Training on fold [4/5]\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "[I 2023-07-29 13:33:30,297] Trial 11 finished with value: 0.7310622231884512 and parameters: {'n_estimators': 1100, 'learning_rate': 0.04499864384752634, 'l2_leaf_reg': 1.1328597564633878, 'max_depth': 10}. Best is trial 10 with value: 0.7333003166074498.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "bestTest = 0.7516508977\n",
            "bestIteration = 1088\n",
            "\n",
            "Training on fold [0/5]\n",
            "\n",
            "bestTest = 0.7286295268\n",
            "bestIteration = 421\n",
            "\n",
            "Training on fold [1/5]\n",
            "\n",
            "bestTest = 0.7262604093\n",
            "bestIteration = 1099\n",
            "\n",
            "Training on fold [2/5]\n",
            "\n",
            "bestTest = 0.7368722161\n",
            "bestIteration = 894\n",
            "\n",
            "Training on fold [3/5]\n",
            "\n",
            "bestTest = 0.7294408745\n",
            "bestIteration = 1078\n",
            "\n",
            "Training on fold [4/5]\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "[I 2023-07-29 13:44:47,158] Trial 12 finished with value: 0.7335444816701011 and parameters: {'n_estimators': 1100, 'learning_rate': 0.05004740324708718, 'l2_leaf_reg': 1.0761494435435939, 'max_depth': 10}. Best is trial 12 with value: 0.7335444816701011.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "bestTest = 0.747424761\n",
            "bestIteration = 1079\n",
            "\n",
            "Training on fold [0/5]\n",
            "\n",
            "bestTest = 0.7358183784\n",
            "bestIteration = 1096\n",
            "\n",
            "Training on fold [1/5]\n",
            "\n",
            "bestTest = 0.7209371571\n",
            "bestIteration = 1097\n",
            "\n",
            "Training on fold [2/5]\n",
            "\n",
            "bestTest = 0.7306210057\n",
            "bestIteration = 1099\n",
            "\n",
            "Training on fold [3/5]\n",
            "\n",
            "bestTest = 0.72896942\n",
            "bestIteration = 1095\n",
            "\n",
            "Training on fold [4/5]\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "[I 2023-07-29 13:56:17,313] Trial 13 finished with value: 0.7301479262648336 and parameters: {'n_estimators': 1100, 'learning_rate': 0.010998904992167786, 'l2_leaf_reg': 4.082643943639809, 'max_depth': 10}. Best is trial 12 with value: 0.7335444816701011.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "bestTest = 0.7347423146\n",
            "bestIteration = 1097\n",
            "\n",
            "Training on fold [0/5]\n",
            "\n",
            "bestTest = 0.7355246595\n",
            "bestIteration = 1099\n",
            "\n",
            "Training on fold [1/5]\n",
            "\n",
            "bestTest = 0.7217384288\n",
            "bestIteration = 1095\n",
            "\n",
            "Training on fold [2/5]\n",
            "\n",
            "bestTest = 0.7282478536\n",
            "bestIteration = 1094\n",
            "\n",
            "Training on fold [3/5]\n",
            "\n",
            "bestTest = 0.7287151253\n",
            "bestIteration = 1074\n",
            "\n",
            "Training on fold [4/5]\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "[I 2023-07-29 14:08:05,184] Trial 14 finished with value: 0.7318923105529512 and parameters: {'n_estimators': 1100, 'learning_rate': 0.012600020988182626, 'l2_leaf_reg': 2.3242614736702736, 'max_depth': 10}. Best is trial 12 with value: 0.7335444816701011.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "bestTest = 0.7454307672\n",
            "bestIteration = 1098\n",
            "\n",
            "Training on fold [0/5]\n",
            "\n",
            "bestTest = 0.7382351043\n",
            "bestIteration = 1036\n",
            "\n",
            "Training on fold [1/5]\n",
            "\n",
            "bestTest = 0.7210678781\n",
            "bestIteration = 910\n",
            "\n",
            "Training on fold [2/5]\n",
            "\n",
            "bestTest = 0.74320896\n",
            "bestIteration = 1040\n",
            "\n",
            "Training on fold [3/5]\n",
            "\n",
            "bestTest = 0.7317222617\n",
            "bestIteration = 1048\n",
            "\n",
            "Training on fold [4/5]\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "[I 2023-07-29 14:19:09,040] Trial 15 finished with value: 0.7341745670180305 and parameters: {'n_estimators': 1050, 'learning_rate': 0.025326582236554013, 'l2_leaf_reg': 3.3411353525643577, 'max_depth': 10}. Best is trial 15 with value: 0.7341745670180305.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "bestTest = 0.7393115474\n",
            "bestIteration = 645\n",
            "\n",
            "Training on fold [0/5]\n",
            "\n",
            "bestTest = 0.7293291266\n",
            "bestIteration = 1049\n",
            "\n",
            "Training on fold [1/5]\n",
            "\n",
            "bestTest = 0.713106804\n",
            "bestIteration = 1049\n",
            "\n",
            "Training on fold [2/5]\n",
            "\n",
            "bestTest = 0.7255366019\n",
            "bestIteration = 1047\n",
            "\n",
            "Training on fold [3/5]\n",
            "\n",
            "bestTest = 0.7224223392\n",
            "bestIteration = 1049\n",
            "\n",
            "Training on fold [4/5]\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "[I 2023-07-29 14:29:58,166] Trial 16 finished with value: 0.7240346475022597 and parameters: {'n_estimators': 1050, 'learning_rate': 0.008554975066641152, 'l2_leaf_reg': 3.761195502183627, 'max_depth': 10}. Best is trial 15 with value: 0.7341745670180305.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "bestTest = 0.7298913394\n",
            "bestIteration = 1048\n",
            "\n",
            "Training on fold [0/5]\n",
            "\n",
            "bestTest = 0.7384707249\n",
            "bestIteration = 961\n",
            "\n",
            "Training on fold [1/5]\n",
            "\n",
            "bestTest = 0.7267598928\n",
            "bestIteration = 1045\n",
            "\n",
            "Training on fold [2/5]\n",
            "\n",
            "bestTest = 0.7316248144\n",
            "bestIteration = 1048\n",
            "\n",
            "Training on fold [3/5]\n",
            "\n",
            "bestTest = 0.7305549277\n",
            "bestIteration = 1008\n",
            "\n",
            "Training on fold [4/5]\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "[I 2023-07-29 14:41:16,704] Trial 17 finished with value: 0.7325488820697733 and parameters: {'n_estimators': 1050, 'learning_rate': 0.02356154820029058, 'l2_leaf_reg': 2.19906926820078, 'max_depth': 10}. Best is trial 15 with value: 0.7341745670180305.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "bestTest = 0.7389119414\n",
            "bestIteration = 776\n",
            "\n",
            "Training on fold [0/5]\n",
            "\n",
            "bestTest = 0.7330377316\n",
            "bestIteration = 886\n",
            "\n",
            "Training on fold [1/5]\n",
            "\n",
            "bestTest = 0.718806081\n",
            "bestIteration = 117\n",
            "\n",
            "Training on fold [2/5]\n",
            "\n",
            "bestTest = 0.7344732425\n",
            "bestIteration = 909\n",
            "\n",
            "Training on fold [3/5]\n",
            "\n",
            "bestTest = 0.7273177151\n",
            "bestIteration = 752\n",
            "\n",
            "Training on fold [4/5]\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "[I 2023-07-29 14:51:03,288] Trial 18 finished with value: 0.7281750189079046 and parameters: {'n_estimators': 1050, 'learning_rate': 0.1073465426635579, 'l2_leaf_reg': 3.2972105642764027, 'max_depth': 10}. Best is trial 15 with value: 0.7341745670180305.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "bestTest = 0.7357611082\n",
            "bestIteration = 1045\n",
            "\n",
            "Training on fold [0/5]\n",
            "\n",
            "bestTest = 0.7230141695\n",
            "bestIteration = 947\n",
            "\n",
            "Training on fold [1/5]\n",
            "\n",
            "bestTest = 0.7088083403\n",
            "bestIteration = 949\n",
            "\n",
            "Training on fold [2/5]\n",
            "\n",
            "bestTest = 0.720583726\n",
            "bestIteration = 948\n",
            "\n",
            "Training on fold [3/5]\n",
            "\n",
            "bestTest = 0.716630877\n",
            "bestIteration = 948\n",
            "\n",
            "Training on fold [4/5]\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "[I 2023-07-29 15:00:48,324] Trial 19 finished with value: 0.7187307374444685 and parameters: {'n_estimators': 950, 'learning_rate': 0.006070258352325702, 'l2_leaf_reg': 4.5739924614595715, 'max_depth': 10}. Best is trial 15 with value: 0.7341745670180305.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "bestTest = 0.7246932317\n",
            "bestIteration = 948\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "best_params_cb = study_cb.best_params\n",
        "# best_params_cb = {'n_estimators': 1050,\n",
        "#                   'learning_rate': 0.025326582236554013,\n",
        "#                   'l2_leaf_reg': 3.3411353525643577,\n",
        "#                   'max_depth': 10}\n",
        "# 0.7341745670180305"
      ],
      "metadata": {
        "id": "uFgMxPfaGbqn"
      },
      "execution_count": 85,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Смотрю, на какой итерации в среднем модель начинает переобучаться:"
      ],
      "metadata": {
        "id": "ZPaB5Cp0O9WS"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "avg_best_est = round(np.mean([645, 1048, 1040, 910, 1036]) + 1)\n",
        "avg_best_est"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "vZdeQpp3oH3N",
        "outputId": "94ad7e3a-3616-4565-af61-1239dd164d5f"
      },
      "execution_count": 87,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "937"
            ]
          },
          "metadata": {},
          "execution_count": 87
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Записываю это ограничение в параметры:"
      ],
      "metadata": {
        "id": "-Q9qK7lXPMmY"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "best_params_cb['n_estimators'] = avg_best_est\n",
        "best_params_cb"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "0FRM8NvlELrx",
        "outputId": "3e983b02-c586-4b2c-e4d3-97f7d9265578"
      },
      "execution_count": 88,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'n_estimators': 937,\n",
              " 'learning_rate': 0.025326582236554013,\n",
              " 'l2_leaf_reg': 3.3411353525643577,\n",
              " 'max_depth': 10}"
            ]
          },
          "metadata": {},
          "execution_count": 88
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Подбор оптимального порога"
      ],
      "metadata": {
        "id": "f6G_3Dk1Pg-Y"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Теперь подберём оптимальный порог для лучшей модели на стратифицированной кросс-валидации на пяти фолдах:"
      ],
      "metadata": {
        "id": "yY00pTaqhTsO"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "cb = CatBoostClassifier(**best_params_cb,\n",
        "                        cat_features=cat_cols,\n",
        "                        verbose=100,\n",
        "                        random_state=42)"
      ],
      "metadata": {
        "id": "q1LiSNiVGdwZ"
      },
      "execution_count": 91,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "kf = StratifiedKFold(n_splits=5, shuffle=True, random_state=42)"
      ],
      "metadata": {
        "id": "0lXhcJklLB8d"
      },
      "execution_count": 92,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# тестируемые пороги\n",
        "threshold_values = np.arange(0.01, 0.8, 0.005)\n",
        "\n",
        "f1_scores = {}\n",
        "\n",
        "for fold, (train_index, val_index) in tqdm(enumerate(kf.split(X, y), start=1)):\n",
        "    X_train, X_val = X.loc[train_index, :], X.loc[val_index, :]\n",
        "    y_train, y_val = y[train_index], y[val_index]\n",
        "\n",
        "    cb.fit(X_train, y_train)\n",
        "\n",
        "    # фиксируемые предсказанные вероятности\n",
        "    y_val_probs = cb.predict_proba(X_val)[:, 1]\n",
        "\n",
        "    for threshold in threshold_values:\n",
        "        # создаём массив предсказаний (0/1) для каждого порога классификации\n",
        "        y_val_pred = (y_val_probs >= threshold).astype(int)\n",
        "        # считаем f1_score для этих предсказаний, складываем в словарь\n",
        "        f1_scores.setdefault(threshold, []).append(f1_score(y_val, y_val_pred))\n",
        "\n",
        "# Считаем среднее значение f1_score для каждого порога\n",
        "average_f1_scores = \\\n",
        "{threshold: np.mean(scores) for threshold, scores in f1_scores.items()}\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "eauCKI6gGiJQ",
        "outputId": "13e8e80a-ea5a-47b6-d06c-683f4ef6e62e"
      },
      "execution_count": 93,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "\r0it [00:00, ?it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "0:\tlearn: 0.6809829\ttotal: 120ms\tremaining: 1m 52s\n",
            "100:\tlearn: 0.4078199\ttotal: 12.2s\tremaining: 1m 41s\n",
            "200:\tlearn: 0.3520537\ttotal: 25.3s\tremaining: 1m 32s\n",
            "300:\tlearn: 0.3141630\ttotal: 37.1s\tremaining: 1m 18s\n",
            "400:\tlearn: 0.2797910\ttotal: 49.6s\tremaining: 1m 6s\n",
            "500:\tlearn: 0.2434173\ttotal: 1m 1s\tremaining: 53.9s\n",
            "600:\tlearn: 0.2122954\ttotal: 1m 14s\tremaining: 41.7s\n",
            "700:\tlearn: 0.1872960\ttotal: 1m 27s\tremaining: 29.4s\n",
            "800:\tlearn: 0.1658436\ttotal: 1m 40s\tremaining: 17s\n",
            "900:\tlearn: 0.1481545\ttotal: 1m 52s\tremaining: 4.51s\n",
            "936:\tlearn: 0.1423833\ttotal: 1m 58s\tremaining: 0us\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "\r1it [01:59, 119.31s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "0:\tlearn: 0.6809803\ttotal: 115ms\tremaining: 1m 47s\n",
            "100:\tlearn: 0.4149719\ttotal: 11.5s\tremaining: 1m 35s\n",
            "200:\tlearn: 0.3507137\ttotal: 23.2s\tremaining: 1m 24s\n",
            "300:\tlearn: 0.3092312\ttotal: 35.6s\tremaining: 1m 15s\n",
            "400:\tlearn: 0.2727434\ttotal: 48s\tremaining: 1m 4s\n",
            "500:\tlearn: 0.2390782\ttotal: 1m\tremaining: 52.9s\n",
            "600:\tlearn: 0.2115478\ttotal: 1m 13s\tremaining: 41.2s\n",
            "700:\tlearn: 0.1879515\ttotal: 1m 26s\tremaining: 29.2s\n",
            "800:\tlearn: 0.1647067\ttotal: 1m 39s\tremaining: 16.9s\n",
            "900:\tlearn: 0.1458862\ttotal: 1m 52s\tremaining: 4.51s\n",
            "936:\tlearn: 0.1389741\ttotal: 1m 57s\tremaining: 0us\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "\r2it [03:57, 118.43s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "0:\tlearn: 0.6811597\ttotal: 130ms\tremaining: 2m 1s\n",
            "100:\tlearn: 0.4117253\ttotal: 11.9s\tremaining: 1m 38s\n",
            "200:\tlearn: 0.3522096\ttotal: 24.8s\tremaining: 1m 30s\n",
            "300:\tlearn: 0.3145793\ttotal: 36.8s\tremaining: 1m 17s\n",
            "400:\tlearn: 0.2760464\ttotal: 49.8s\tremaining: 1m 6s\n",
            "500:\tlearn: 0.2423007\ttotal: 1m 2s\tremaining: 54.7s\n",
            "600:\tlearn: 0.2132452\ttotal: 1m 15s\tremaining: 42.3s\n",
            "700:\tlearn: 0.1894680\ttotal: 1m 28s\tremaining: 29.8s\n",
            "800:\tlearn: 0.1680300\ttotal: 1m 41s\tremaining: 17.2s\n",
            "900:\tlearn: 0.1490316\ttotal: 1m 53s\tremaining: 4.54s\n",
            "936:\tlearn: 0.1430855\ttotal: 1m 58s\tremaining: 0us\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "\r3it [05:56, 118.89s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "0:\tlearn: 0.6813114\ttotal: 40.6ms\tremaining: 38s\n",
            "100:\tlearn: 0.4145112\ttotal: 11.1s\tremaining: 1m 32s\n",
            "200:\tlearn: 0.3542441\ttotal: 24s\tremaining: 1m 27s\n",
            "300:\tlearn: 0.3159169\ttotal: 34.6s\tremaining: 1m 13s\n",
            "400:\tlearn: 0.2778268\ttotal: 47s\tremaining: 1m 2s\n",
            "500:\tlearn: 0.2428819\ttotal: 59.9s\tremaining: 52.2s\n",
            "600:\tlearn: 0.2132648\ttotal: 1m 12s\tremaining: 40.8s\n",
            "700:\tlearn: 0.1890572\ttotal: 1m 26s\tremaining: 29s\n",
            "800:\tlearn: 0.1698219\ttotal: 1m 39s\tremaining: 16.8s\n",
            "900:\tlearn: 0.1511073\ttotal: 1m 52s\tremaining: 4.48s\n",
            "936:\tlearn: 0.1442161\ttotal: 1m 56s\tremaining: 0us\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "\r4it [07:54, 118.40s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "0:\tlearn: 0.6813531\ttotal: 79.6ms\tremaining: 1m 14s\n",
            "100:\tlearn: 0.4177914\ttotal: 10.5s\tremaining: 1m 27s\n",
            "200:\tlearn: 0.3523547\ttotal: 23s\tremaining: 1m 24s\n",
            "300:\tlearn: 0.3133053\ttotal: 34.8s\tremaining: 1m 13s\n",
            "400:\tlearn: 0.2755508\ttotal: 47.2s\tremaining: 1m 3s\n",
            "500:\tlearn: 0.2401058\ttotal: 59.9s\tremaining: 52.2s\n",
            "600:\tlearn: 0.2095403\ttotal: 1m 13s\tremaining: 40.8s\n",
            "700:\tlearn: 0.1867883\ttotal: 1m 25s\tremaining: 28.9s\n",
            "800:\tlearn: 0.1665625\ttotal: 1m 38s\tremaining: 16.8s\n",
            "900:\tlearn: 0.1471482\ttotal: 1m 51s\tremaining: 4.46s\n",
            "936:\tlearn: 0.1410588\ttotal: 1m 56s\tremaining: 0us\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "5it [09:51, 118.22s/it]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Сортируем словарь по значениям *F1*-скора по убыванию:"
      ],
      "metadata": {
        "id": "Q1sHXPMzjEpX"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "best_scores_cb = sorted(average_f1_scores.items(), key=lambda x: -x[1])\n",
        "best_scores_cb[:10]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "W3nHkY6_GlER",
        "outputId": "43139282-3c6d-4310-cab0-c0d15604d2af"
      },
      "execution_count": 94,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[(0.19999999999999998, 0.4559088033391901),\n",
              " (0.19499999999999998, 0.45539804255328536),\n",
              " (0.18999999999999997, 0.4540894510904253),\n",
              " (0.18499999999999997, 0.45395310164452746),\n",
              " (0.18, 0.45340580656836915),\n",
              " (0.15999999999999998, 0.45307127385183305),\n",
              " (0.16499999999999998, 0.4528876516632181),\n",
              " (0.205, 0.4522133894185247),\n",
              " (0.20999999999999996, 0.4517636619216856),\n",
              " (0.16999999999999998, 0.45131808029458104)]"
            ]
          },
          "metadata": {},
          "execution_count": 94
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "best_thresh_cb = round(best_scores_cb[0][0], 5)"
      ],
      "metadata": {
        "id": "2tfBjGyRjjrI"
      },
      "execution_count": 95,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Оптимизация случайного леса"
      ],
      "metadata": {
        "id": "b8EEieJ6M02F"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Лучшие гиперпараметры случайного леса также ищу на стратифицированной кросс-валидации на пяти фолдах. Напоминаю, что матрица признаков для случайного леса — $100$ наиболее коллериующих с целевым признаком непрерывных исходных и интерактивных фичей, а также *OHE*-закодированные категориальные признаки."
      ],
      "metadata": {
        "id": "N-5D-xKTUBcR"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "X_rfc.shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "C0e2Ja45UTDc",
        "outputId": "a9cfeaa3-ef3d-4257-c579-8f0fc5f9d41b"
      },
      "execution_count": 96,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(13863, 111)"
            ]
          },
          "metadata": {},
          "execution_count": 96
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Подбор гиперпараметров"
      ],
      "metadata": {
        "id": "w5ATpGITM02U"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def objective_rfc(n_trials):\n",
        "\n",
        "    params = {\n",
        "        'criterion':\n",
        "        n_trials.suggest_categorical('criterion',\n",
        "                                     ['gini', 'entropy']),\n",
        "        'max_depth':\n",
        "        n_trials.suggest_int('max_depth', 12, 17),\n",
        "        'n_estimators':\n",
        "        n_trials.suggest_int('n_estimators', 500, 1000),\n",
        "        'random_state':\n",
        "        42\n",
        "    }\n",
        "\n",
        "    rfc = RandomForestClassifier(**params)\n",
        "    roc_auc = np.mean(\n",
        "        cross_val_score(rfc,\n",
        "                        X_rfc,\n",
        "                        y,\n",
        "                        scoring='roc_auc',\n",
        "                        cv=kf))\n",
        "\n",
        "    return roc_auc"
      ],
      "metadata": {
        "id": "5GOhm6LzTgaP"
      },
      "execution_count": 97,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "study_rfc = optuna.create_study(\n",
        "    direction='maximize',\n",
        "    pruner=HyperbandPruner(),\n",
        "    sampler=optuna.samplers.TPESampler(seed=42))\n",
        "study_rfc.optimize(objective_rfc, n_trials=20)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "62b9926e-27cf-4eb8-f4f5-a7fbcc75f952",
        "id": "t8U9emGbTgaP"
      },
      "execution_count": 98,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "[I 2023-07-29 15:13:44,796] A new study created in memory with name: no-name-1ecc6180-0f3c-4ba3-8fba-cb02a88e3e9b\n",
            "[I 2023-07-29 15:19:57,568] Trial 0 finished with value: 0.7351714186578381 and parameters: {'criterion': 'entropy', 'max_depth': 16, 'n_estimators': 799}. Best is trial 0 with value: 0.7351714186578381.\n",
            "[I 2023-07-29 15:25:07,149] Trial 1 finished with value: 0.7135919251457097 and parameters: {'criterion': 'gini', 'max_depth': 12, 'n_estimators': 933}. Best is trial 0 with value: 0.7351714186578381.\n",
            "[I 2023-07-29 15:31:33,747] Trial 2 finished with value: 0.718992140612358 and parameters: {'criterion': 'entropy', 'max_depth': 12, 'n_estimators': 985}. Best is trial 0 with value: 0.7351714186578381.\n",
            "[I 2023-07-29 15:35:02,822] Trial 3 finished with value: 0.7170538418854417 and parameters: {'criterion': 'gini', 'max_depth': 13, 'n_estimators': 591}. Best is trial 0 with value: 0.7351714186578381.\n",
            "[I 2023-07-29 15:39:43,038] Trial 4 finished with value: 0.7288984406217084 and parameters: {'criterion': 'entropy', 'max_depth': 14, 'n_estimators': 645}. Best is trial 0 with value: 0.7351714186578381.\n",
            "[I 2023-07-29 15:43:43,473] Trial 5 finished with value: 0.7168627733358097 and parameters: {'criterion': 'gini', 'max_depth': 13, 'n_estimators': 683}. Best is trial 0 with value: 0.7351714186578381.\n",
            "[I 2023-07-29 15:48:57,087] Trial 6 finished with value: 0.7229618276137239 and parameters: {'criterion': 'entropy', 'max_depth': 13, 'n_estimators': 757}. Best is trial 0 with value: 0.7351714186578381.\n",
            "[I 2023-07-29 15:52:44,497] Trial 7 finished with value: 0.7233327892337417 and parameters: {'criterion': 'gini', 'max_depth': 15, 'n_estimators': 585}. Best is trial 0 with value: 0.7351714186578381.\n",
            "[I 2023-07-29 15:59:56,563] Trial 8 finished with value: 0.7391198854103351 and parameters: {'criterion': 'entropy', 'max_depth': 17, 'n_estimators': 905}. Best is trial 8 with value: 0.7391198854103351.\n",
            "[I 2023-07-29 16:04:47,280] Trial 9 finished with value: 0.7268608604389435 and parameters: {'criterion': 'gini', 'max_depth': 16, 'n_estimators': 720}. Best is trial 8 with value: 0.7391198854103351.\n",
            "[I 2023-07-29 16:11:31,717] Trial 10 finished with value: 0.7389537766664491 and parameters: {'criterion': 'entropy', 'max_depth': 17, 'n_estimators': 849}. Best is trial 8 with value: 0.7391198854103351.\n",
            "[I 2023-07-29 16:18:22,830] Trial 11 finished with value: 0.7390309388067948 and parameters: {'criterion': 'entropy', 'max_depth': 17, 'n_estimators': 861}. Best is trial 8 with value: 0.7391198854103351.\n",
            "[I 2023-07-29 16:25:27,762] Trial 12 finished with value: 0.7389950967417002 and parameters: {'criterion': 'entropy', 'max_depth': 17, 'n_estimators': 888}. Best is trial 8 with value: 0.7391198854103351.\n",
            "[I 2023-07-29 16:31:59,185] Trial 13 finished with value: 0.7354057653408506 and parameters: {'criterion': 'entropy', 'max_depth': 16, 'n_estimators': 841}. Best is trial 8 with value: 0.7391198854103351.\n",
            "[I 2023-07-29 16:39:56,597] Trial 14 finished with value: 0.7393029138689057 and parameters: {'criterion': 'entropy', 'max_depth': 17, 'n_estimators': 997}. Best is trial 14 with value: 0.7393029138689057.\n",
            "[I 2023-07-29 16:47:27,637] Trial 15 finished with value: 0.733089690531386 and parameters: {'criterion': 'entropy', 'max_depth': 15, 'n_estimators': 1000}. Best is trial 14 with value: 0.7393029138689057.\n",
            "[I 2023-07-29 16:54:40,985] Trial 16 finished with value: 0.7354014149031058 and parameters: {'criterion': 'entropy', 'max_depth': 16, 'n_estimators': 932}. Best is trial 14 with value: 0.7393029138689057.\n",
            "[I 2023-07-29 17:02:08,333] Trial 17 finished with value: 0.7391048745798032 and parameters: {'criterion': 'entropy', 'max_depth': 17, 'n_estimators': 936}. Best is trial 14 with value: 0.7393029138689057.\n",
            "[I 2023-07-29 17:08:07,460] Trial 18 finished with value: 0.7326832332528611 and parameters: {'criterion': 'entropy', 'max_depth': 15, 'n_estimators': 795}. Best is trial 14 with value: 0.7393029138689057.\n",
            "[I 2023-07-29 17:12:20,978] Trial 19 finished with value: 0.7382827508459384 and parameters: {'criterion': 'entropy', 'max_depth': 17, 'n_estimators': 532}. Best is trial 14 with value: 0.7393029138689057.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "best_params_rfc = study_rfc.best_params\n",
        "print(study_rfc.best_value)\n",
        "print(study_rfc.best_params)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "eace1d1c-883e-4770-a49f-d17d5f47fcd9",
        "id": "yX71GciJTgaQ"
      },
      "execution_count": 99,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "0.7393029138689057\n",
            "{'criterion': 'entropy', 'max_depth': 17, 'n_estimators': 997}\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Подбор оптимального порога"
      ],
      "metadata": {
        "id": "0MkGulP7M02V"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "rfc = RandomForestClassifier(**best_params_rfc, random_state=42)"
      ],
      "metadata": {
        "id": "0S0kkhgNS4n_"
      },
      "execution_count": 100,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# тестируемые пороги\n",
        "threshold_values = np.arange(0, 1, 0.005)\n",
        "\n",
        "f1_scores = {}\n",
        "\n",
        "\n",
        "for fold, (train_index, val_index) in tqdm(enumerate(kf.split(X_rfc, y),\n",
        "                                                     start=1)):\n",
        "    X_train, X_val = X_rfc.loc[train_index, :], X_rfc.loc[val_index, :]\n",
        "    y_train, y_val = y[train_index], y[val_index]\n",
        "\n",
        "    rfc.fit(X_train, y_train)\n",
        "\n",
        "    # фиксируемые предсказанные вероятности\n",
        "    y_val_probs = rfc.predict_proba(X_val)[:, 1]\n",
        "\n",
        "    for threshold in threshold_values:\n",
        "        # создаём массив предсказаний (0/1) для каждого порога классификации\n",
        "        y_val_pred = (y_val_probs >= threshold).astype(int)\n",
        "        # считаем f1_score для этих предсказаний, складываем в словарь\n",
        "        f1_scores.setdefault(threshold, []).append(f1_score(y_val, y_val_pred))\n",
        "\n",
        "# Считаем среднее значение f1_score для каждого порога\n",
        "average_f1_scores = \\\n",
        " {threshold: np.mean(scores) for threshold, scores in f1_scores.items()}\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "723d81b1-f808-43c3-fd8d-8e5407421d94",
        "id": "dlsabh_sM02W"
      },
      "execution_count": 101,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "5it [08:02, 96.48s/it]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "best_scores_rfc = sorted(average_f1_scores.items(), key=lambda x: -x[1])\n",
        "best_scores_rfc[:10]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "a448c746-13a5-4d93-bbc2-fcfec75b0f53",
        "id": "OmckKrYIM02X"
      },
      "execution_count": 102,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[(0.23, 0.4564382028463985),\n",
              " (0.21, 0.4553706543277373),\n",
              " (0.22, 0.45531512646897376),\n",
              " (0.225, 0.4549352320107225),\n",
              " (0.2, 0.4541292595001482),\n",
              " (0.20500000000000002, 0.45408374931745055),\n",
              " (0.195, 0.45383244343860313),\n",
              " (0.23500000000000001, 0.45378557283034204),\n",
              " (0.215, 0.45321461930779405),\n",
              " (0.19, 0.4524036490320277)]"
            ]
          },
          "metadata": {},
          "execution_count": 102
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "best_thresh_rfc = round(best_scores_rfc[0][0], 5)"
      ],
      "metadata": {
        "id": "zpU58fYMM02X"
      },
      "execution_count": 103,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Блендинг двух моделей"
      ],
      "metadata": {
        "id": "LAUWkuSOBJY2"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "best_thresh_cb, best_thresh_rfc"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "naMQFVUIGont",
        "outputId": "6ca55116-dc86-4ed8-973f-4a2ddc3c874b"
      },
      "execution_count": 104,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(0.2, 0.23)"
            ]
          },
          "metadata": {},
          "execution_count": 104
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Обучаем *CatBoost* на всей тренировочной выборке:"
      ],
      "metadata": {
        "id": "PuVd3p-RmtSf"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "cb.fit(X, y)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "eO7W6lRWGriq",
        "outputId": "e2527459-b0a0-4053-f24c-73910116df14"
      },
      "execution_count": 105,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "0:\tlearn: 0.6809927\ttotal: 208ms\tremaining: 3m 14s\n",
            "100:\tlearn: 0.4191416\ttotal: 13.2s\tremaining: 1m 49s\n",
            "200:\tlearn: 0.3635794\ttotal: 26.6s\tremaining: 1m 37s\n",
            "300:\tlearn: 0.3259959\ttotal: 39.8s\tremaining: 1m 24s\n",
            "400:\tlearn: 0.2953296\ttotal: 52.7s\tremaining: 1m 10s\n",
            "500:\tlearn: 0.2632193\ttotal: 1m 6s\tremaining: 57.9s\n",
            "600:\tlearn: 0.2360861\ttotal: 1m 20s\tremaining: 45s\n",
            "700:\tlearn: 0.2124898\ttotal: 1m 34s\tremaining: 31.8s\n",
            "800:\tlearn: 0.1898069\ttotal: 1m 48s\tremaining: 18.4s\n",
            "900:\tlearn: 0.1701735\ttotal: 2m 2s\tremaining: 4.91s\n",
            "936:\tlearn: 0.1644982\ttotal: 2m 7s\tremaining: 0us\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<catboost.core.CatBoostClassifier at 0x7f3bb015fa30>"
            ]
          },
          "metadata": {},
          "execution_count": 105
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Записываем предсказания вероятностей и предсказания классов в переменные:"
      ],
      "metadata": {
        "id": "TvdcQvhIm1-x"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "cb_probas = cb.predict_proba(test)[:, 1]\n",
        "cb_just_preds = cb.predict(test)\n",
        "cb_probas[:10]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "PrUCpLFNBb-W",
        "outputId": "201acc88-2281-45b4-8fd9-9e1f946f8d6c"
      },
      "execution_count": 106,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([0.14353557, 0.01332447, 0.15809758, 0.08562579, 0.14607352,\n",
              "       0.0384107 , 0.18213732, 0.07025704, 0.28239483, 0.03873094])"
            ]
          },
          "metadata": {},
          "execution_count": 106
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Повторяем со случайным лесом:"
      ],
      "metadata": {
        "id": "hpzfhlHLm-kp"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "rfc.fit(X_rfc, y)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 91
        },
        "id": "9NWPF_TSEBCB",
        "outputId": "752f7cb8-1d55-4d70-e15c-2f7808b37974"
      },
      "execution_count": 107,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "RandomForestClassifier(criterion='entropy', max_depth=17, n_estimators=997,\n",
              "                       random_state=42)"
            ],
            "text/html": [
              "<style>#sk-container-id-2 {color: black;background-color: white;}#sk-container-id-2 pre{padding: 0;}#sk-container-id-2 div.sk-toggleable {background-color: white;}#sk-container-id-2 label.sk-toggleable__label {cursor: pointer;display: block;width: 100%;margin-bottom: 0;padding: 0.3em;box-sizing: border-box;text-align: center;}#sk-container-id-2 label.sk-toggleable__label-arrow:before {content: \"▸\";float: left;margin-right: 0.25em;color: #696969;}#sk-container-id-2 label.sk-toggleable__label-arrow:hover:before {color: black;}#sk-container-id-2 div.sk-estimator:hover label.sk-toggleable__label-arrow:before {color: black;}#sk-container-id-2 div.sk-toggleable__content {max-height: 0;max-width: 0;overflow: hidden;text-align: left;background-color: #f0f8ff;}#sk-container-id-2 div.sk-toggleable__content pre {margin: 0.2em;color: black;border-radius: 0.25em;background-color: #f0f8ff;}#sk-container-id-2 input.sk-toggleable__control:checked~div.sk-toggleable__content {max-height: 200px;max-width: 100%;overflow: auto;}#sk-container-id-2 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {content: \"▾\";}#sk-container-id-2 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-2 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-2 input.sk-hidden--visually {border: 0;clip: rect(1px 1px 1px 1px);clip: rect(1px, 1px, 1px, 1px);height: 1px;margin: -1px;overflow: hidden;padding: 0;position: absolute;width: 1px;}#sk-container-id-2 div.sk-estimator {font-family: monospace;background-color: #f0f8ff;border: 1px dotted black;border-radius: 0.25em;box-sizing: border-box;margin-bottom: 0.5em;}#sk-container-id-2 div.sk-estimator:hover {background-color: #d4ebff;}#sk-container-id-2 div.sk-parallel-item::after {content: \"\";width: 100%;border-bottom: 1px solid gray;flex-grow: 1;}#sk-container-id-2 div.sk-label:hover label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-2 div.sk-serial::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: 0;}#sk-container-id-2 div.sk-serial {display: flex;flex-direction: column;align-items: center;background-color: white;padding-right: 0.2em;padding-left: 0.2em;position: relative;}#sk-container-id-2 div.sk-item {position: relative;z-index: 1;}#sk-container-id-2 div.sk-parallel {display: flex;align-items: stretch;justify-content: center;background-color: white;position: relative;}#sk-container-id-2 div.sk-item::before, #sk-container-id-2 div.sk-parallel-item::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: -1;}#sk-container-id-2 div.sk-parallel-item {display: flex;flex-direction: column;z-index: 1;position: relative;background-color: white;}#sk-container-id-2 div.sk-parallel-item:first-child::after {align-self: flex-end;width: 50%;}#sk-container-id-2 div.sk-parallel-item:last-child::after {align-self: flex-start;width: 50%;}#sk-container-id-2 div.sk-parallel-item:only-child::after {width: 0;}#sk-container-id-2 div.sk-dashed-wrapped {border: 1px dashed gray;margin: 0 0.4em 0.5em 0.4em;box-sizing: border-box;padding-bottom: 0.4em;background-color: white;}#sk-container-id-2 div.sk-label label {font-family: monospace;font-weight: bold;display: inline-block;line-height: 1.2em;}#sk-container-id-2 div.sk-label-container {text-align: center;}#sk-container-id-2 div.sk-container {/* jupyter's `normalize.less` sets `[hidden] { display: none; }` but bootstrap.min.css set `[hidden] { display: none !important; }` so we also need the `!important` here to be able to override the default hidden behavior on the sphinx rendered scikit-learn.org. See: https://github.com/scikit-learn/scikit-learn/issues/21755 */display: inline-block !important;position: relative;}#sk-container-id-2 div.sk-text-repr-fallback {display: none;}</style><div id=\"sk-container-id-2\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>RandomForestClassifier(criterion=&#x27;entropy&#x27;, max_depth=17, n_estimators=997,\n",
              "                       random_state=42)</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-2\" type=\"checkbox\" checked><label for=\"sk-estimator-id-2\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">RandomForestClassifier</label><div class=\"sk-toggleable__content\"><pre>RandomForestClassifier(criterion=&#x27;entropy&#x27;, max_depth=17, n_estimators=997,\n",
              "                       random_state=42)</pre></div></div></div></div></div>"
            ]
          },
          "metadata": {},
          "execution_count": 107
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "rfc_just_preds = rfc.predict(test_rfc)\n",
        "rfc_probas = rfc.predict_proba(test_rfc)[:, 1]\n",
        "rfc_probas[:10]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "dEJzVfMFEUwQ",
        "outputId": "17b9f4ee-4fb2-4772-adac-8818b0fa4a19"
      },
      "execution_count": 108,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([0.2658456 , 0.05816638, 0.234725  , 0.10524756, 0.22698235,\n",
              "       0.34884203, 0.18219176, 0.15352444, 0.2775249 , 0.10941587])"
            ]
          },
          "metadata": {},
          "execution_count": 108
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Записываем предсказания вероятностей от двух моделей в датафрейм:"
      ],
      "metadata": {
        "id": "SIcGwDS3nGXf"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "two_probas = pd.DataFrame({'id': test_ids,\n",
        "                          'cb_proba': cb_probas,\n",
        "                          'rfc_proba': rfc_probas})\n",
        "two_probas.head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "id": "DArrgi2LE3Va",
        "outputId": "76eae391-4aff-4a09-f2b5-1d94c1900094"
      },
      "execution_count": 109,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "      id  cb_proba  rfc_proba\n",
              "0  19200  0.143536   0.265846\n",
              "1   7784  0.013324   0.058166\n",
              "2  11682  0.158098   0.234725\n",
              "3   9949  0.085626   0.105248\n",
              "4   9378  0.146074   0.226982"
            ],
            "text/html": [
              "\n",
              "\n",
              "  <div id=\"df-2830e854-5652-4763-ac61-f3f3157b9056\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>id</th>\n",
              "      <th>cb_proba</th>\n",
              "      <th>rfc_proba</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>19200</td>\n",
              "      <td>0.143536</td>\n",
              "      <td>0.265846</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>7784</td>\n",
              "      <td>0.013324</td>\n",
              "      <td>0.058166</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>11682</td>\n",
              "      <td>0.158098</td>\n",
              "      <td>0.234725</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>9949</td>\n",
              "      <td>0.085626</td>\n",
              "      <td>0.105248</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>9378</td>\n",
              "      <td>0.146074</td>\n",
              "      <td>0.226982</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-2830e854-5652-4763-ac61-f3f3157b9056')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "\n",
              "\n",
              "\n",
              "    <div id=\"df-9504c6e6-73b3-481a-bbab-8ceced3bc084\">\n",
              "      <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-9504c6e6-73b3-481a-bbab-8ceced3bc084')\"\n",
              "              title=\"Suggest charts.\"\n",
              "              style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "      </button>\n",
              "    </div>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "    background-color: #E8F0FE;\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: #1967D2;\n",
              "    height: 32px;\n",
              "    padding: 0 0 0 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: #E2EBFA;\n",
              "    box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: #174EA6;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "    background-color: #3B4455;\n",
              "    fill: #D2E3FC;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart:hover {\n",
              "    background-color: #434B5C;\n",
              "    box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "    filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "    fill: #FFFFFF;\n",
              "  }\n",
              "</style>\n",
              "\n",
              "    <script>\n",
              "      async function quickchart(key) {\n",
              "        const containerElement = document.querySelector('#' + key);\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      }\n",
              "    </script>\n",
              "\n",
              "      <script>\n",
              "\n",
              "function displayQuickchartButton(domScope) {\n",
              "  let quickchartButtonEl =\n",
              "    domScope.querySelector('#df-9504c6e6-73b3-481a-bbab-8ceced3bc084 button.colab-df-quickchart');\n",
              "  quickchartButtonEl.style.display =\n",
              "    google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "}\n",
              "\n",
              "        displayQuickchartButton(document);\n",
              "      </script>\n",
              "      <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-2830e854-5652-4763-ac61-f3f3157b9056 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-2830e854-5652-4763-ac61-f3f3157b9056');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 109
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Посмотрим, во скольких процентах предсказаний модели несогласны между собой.\n",
        "\n",
        "На объектах где они согласны записываю метку класса ($0/1$), где нет — пару предсказаний вероятностей."
      ],
      "metadata": {
        "id": "4m5pfK0zn0G-"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "two_probas['meld'] = two_probas.apply(\n",
        "    lambda x: 1 if x['cb_proba'] >= best_thresh_cb\n",
        "                   and x['rfc_proba'] >= best_thresh_rfc\n",
        "    else 0 if x['cb_proba'] < best_thresh_cb\n",
        "              and x['rfc_proba'] < best_thresh_rfc\n",
        "    else (x['cb_proba'], x['rfc_proba']),\n",
        "    axis='columns')"
      ],
      "metadata": {
        "id": "4kCyWkg8FLid"
      },
      "execution_count": 110,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "display(two_probas.meld.head())\n",
        "print(f'В {np.mean(cb_just_preds != rfc_just_preds):.2%} \\\n",
        "процентах объектов тестовой выборки модели не согласны')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 139
        },
        "id": "vK6BANfyoBwv",
        "outputId": "f05b78bf-1edd-4281-c919-04b25eea309a"
      },
      "execution_count": 111,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "0    (0.14353556628842706, 0.2658455998472263)\n",
              "1                                            0\n",
              "2    (0.1580975793250085, 0.23472500112887662)\n",
              "3                                            0\n",
              "4                                            0\n",
              "Name: meld, dtype: object"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "В 2.68% процентах объектов тестовой выборки модели не согласны\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Теперь самое интересное: как \"подружить\" разногласия.  \n",
        "\n",
        "Идея простая — дать предпочтение более уверенной модели. \"Уверенность\" буду измерять стандартными отклонениями (*z-score*) от оптимального порога."
      ],
      "metadata": {
        "id": "-nTMfMH3o5UX"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "st_scaler = StandardScaler()"
      ],
      "metadata": {
        "id": "bJ_yeA2kHuGX"
      },
      "execution_count": 112,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Нахожу стандартные отклонения вероятностных предсказаний:"
      ],
      "metadata": {
        "id": "JhKF_dYNYHUm"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "st_scaler.fit(two_probas[['cb_proba', 'rfc_proba']])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 74
        },
        "id": "Fvfe6BcdIFty",
        "outputId": "338e61fd-9f6f-4b62-cf08-92b71f83659b"
      },
      "execution_count": 113,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "StandardScaler()"
            ],
            "text/html": [
              "<style>#sk-container-id-3 {color: black;background-color: white;}#sk-container-id-3 pre{padding: 0;}#sk-container-id-3 div.sk-toggleable {background-color: white;}#sk-container-id-3 label.sk-toggleable__label {cursor: pointer;display: block;width: 100%;margin-bottom: 0;padding: 0.3em;box-sizing: border-box;text-align: center;}#sk-container-id-3 label.sk-toggleable__label-arrow:before {content: \"▸\";float: left;margin-right: 0.25em;color: #696969;}#sk-container-id-3 label.sk-toggleable__label-arrow:hover:before {color: black;}#sk-container-id-3 div.sk-estimator:hover label.sk-toggleable__label-arrow:before {color: black;}#sk-container-id-3 div.sk-toggleable__content {max-height: 0;max-width: 0;overflow: hidden;text-align: left;background-color: #f0f8ff;}#sk-container-id-3 div.sk-toggleable__content pre {margin: 0.2em;color: black;border-radius: 0.25em;background-color: #f0f8ff;}#sk-container-id-3 input.sk-toggleable__control:checked~div.sk-toggleable__content {max-height: 200px;max-width: 100%;overflow: auto;}#sk-container-id-3 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {content: \"▾\";}#sk-container-id-3 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-3 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-3 input.sk-hidden--visually {border: 0;clip: rect(1px 1px 1px 1px);clip: rect(1px, 1px, 1px, 1px);height: 1px;margin: -1px;overflow: hidden;padding: 0;position: absolute;width: 1px;}#sk-container-id-3 div.sk-estimator {font-family: monospace;background-color: #f0f8ff;border: 1px dotted black;border-radius: 0.25em;box-sizing: border-box;margin-bottom: 0.5em;}#sk-container-id-3 div.sk-estimator:hover {background-color: #d4ebff;}#sk-container-id-3 div.sk-parallel-item::after {content: \"\";width: 100%;border-bottom: 1px solid gray;flex-grow: 1;}#sk-container-id-3 div.sk-label:hover label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-3 div.sk-serial::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: 0;}#sk-container-id-3 div.sk-serial {display: flex;flex-direction: column;align-items: center;background-color: white;padding-right: 0.2em;padding-left: 0.2em;position: relative;}#sk-container-id-3 div.sk-item {position: relative;z-index: 1;}#sk-container-id-3 div.sk-parallel {display: flex;align-items: stretch;justify-content: center;background-color: white;position: relative;}#sk-container-id-3 div.sk-item::before, #sk-container-id-3 div.sk-parallel-item::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: -1;}#sk-container-id-3 div.sk-parallel-item {display: flex;flex-direction: column;z-index: 1;position: relative;background-color: white;}#sk-container-id-3 div.sk-parallel-item:first-child::after {align-self: flex-end;width: 50%;}#sk-container-id-3 div.sk-parallel-item:last-child::after {align-self: flex-start;width: 50%;}#sk-container-id-3 div.sk-parallel-item:only-child::after {width: 0;}#sk-container-id-3 div.sk-dashed-wrapped {border: 1px dashed gray;margin: 0 0.4em 0.5em 0.4em;box-sizing: border-box;padding-bottom: 0.4em;background-color: white;}#sk-container-id-3 div.sk-label label {font-family: monospace;font-weight: bold;display: inline-block;line-height: 1.2em;}#sk-container-id-3 div.sk-label-container {text-align: center;}#sk-container-id-3 div.sk-container {/* jupyter's `normalize.less` sets `[hidden] { display: none; }` but bootstrap.min.css set `[hidden] { display: none !important; }` so we also need the `!important` here to be able to override the default hidden behavior on the sphinx rendered scikit-learn.org. See: https://github.com/scikit-learn/scikit-learn/issues/21755 */display: inline-block !important;position: relative;}#sk-container-id-3 div.sk-text-repr-fallback {display: none;}</style><div id=\"sk-container-id-3\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>StandardScaler()</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-3\" type=\"checkbox\" checked><label for=\"sk-estimator-id-3\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">StandardScaler</label><div class=\"sk-toggleable__content\"><pre>StandardScaler()</pre></div></div></div></div></div>"
            ]
          },
          "metadata": {},
          "execution_count": 113
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "print(st_scaler.scale_) # , st_scaler.mean_\n",
        "std_cb = st_scaler.scale_[0]\n",
        "std_rfc = st_scaler.scale_[1]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "GgdF-mFvIaQp",
        "outputId": "3ff917eb-6c05-443d-f40c-aac86332a63b"
      },
      "execution_count": 114,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[0.13808129 0.11503701]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Применяю такую стандартизацию к значениям внутри пар по разные стороны от соответствующих оптимальных порогов:"
      ],
      "metadata": {
        "id": "iFOg-p2Mp_sI"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "two_probas['z_transform'] = two_probas['meld'].apply(\n",
        "    lambda x: x if isinstance(x, int)\n",
        "    else ((x[0] - best_thresh_cb) / std_cb,\n",
        "          (x[1] - best_thresh_rfc) / std_rfc)\n",
        ")"
      ],
      "metadata": {
        "id": "Tbt7OCKkIMHf"
      },
      "execution_count": 115,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Теперь напишу функцию для конвертации несогласий в бинарные предсказания. Аргумент `coef` будет означать относительный вес предсказаний *CatBoost*-а."
      ],
      "metadata": {
        "id": "2mLsWJfYyVQb"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def decide(x: Union[int, tuple], coef: Union[int, float] = 1) -> int:\n",
        "    '''\n",
        "    Функция для конвертации несогласий в бинарные предсказания\n",
        "    с помощью взвешенного сравнения уверенностей\n",
        "    '''\n",
        "    if isinstance(x, int):\n",
        "        return x\n",
        "    if abs(x[0]) * coef >= abs(x[1]):\n",
        "        if x[0] > 0:\n",
        "            return 1\n",
        "        else:\n",
        "            return 0\n",
        "    if abs(x[0]) * coef < abs(x[1]):\n",
        "        if x[1] > 0:\n",
        "            return 1\n",
        "        else:\n",
        "            return 0"
      ],
      "metadata": {
        "id": "Ndb7iuFWPElY"
      },
      "execution_count": 116,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Последний этап обучения можно уже проводить на публичной тестовой выборке ;-)\n",
        "\n",
        "Попробовав несколько коэффициентов, лучшим оказался $0.5$.\n",
        "\n",
        "(По-хорошему, конечно, лучше было бы подбирать этот коэффициент как гиперпараметр на кросс-валидации на тренировочной выборке)"
      ],
      "metadata": {
        "id": "PNOX1XWO0Sot"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "two_probas['equal_weight'] = \\\n",
        "two_probas['z_transform'].apply(lambda x: decide(x, coef=1))"
      ],
      "metadata": {
        "id": "uBDSQUyIP1PX"
      },
      "execution_count": 117,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "two_probas['0_5'] = \\\n",
        "two_probas['z_transform'].apply(lambda x: decide(x, coef=0.5))"
      ],
      "metadata": {
        "id": "toeodh0vQQ1o"
      },
      "execution_count": 118,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "(two_probas['0_5'] != two_probas['equal_weight']).sum()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "5ul2Ks2GQYvv",
        "outputId": "fba80622-fb13-4ac0-de12-8f45e76013cb"
      },
      "execution_count": 119,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "150"
            ]
          },
          "metadata": {},
          "execution_count": 119
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "sub = pd.DataFrame({'id': test_ids, 'smoking': two_probas['0_5']})"
      ],
      "metadata": {
        "id": "v47T-L6yG0z5"
      },
      "execution_count": 120,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "sub.to_csv('0_5_blend.csv', index=False)"
      ],
      "metadata": {
        "id": "j4g9cvXdG3QK"
      },
      "execution_count": 121,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Итоговый скор: **0.479**."
      ],
      "metadata": {
        "id": "149XlQRIcEln"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Спасибо за внимание :)"
      ],
      "metadata": {
        "id": "dzjKbXKqUSJa"
      }
    }
  ]
}