scikit.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "scikit.ipynb",
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9lCWEhqBRvrQ",
        "colab_type": "text"
      },
      "source": [
        "# Topic 1  Overview of Machine Learning and Scikit Learn"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lxIv8Lw2RCa7",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import sklearn"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "mYQ6_65PR7AM",
        "colab_type": "text"
      },
      "source": [
        "# Topic 2 Classification"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qVReRW7FSBqm",
        "colab_type": "text"
      },
      "source": [
        "## Step 1: Prepare the Data"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "FIOjxDL5ZD43",
        "colab_type": "text"
      },
      "source": [
        "### Import Data"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5drY2ePVRu2J",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Load Iris dataset\n",
        "import pandas as pd\n",
        "dataset_path = \"https://raw.githubusercontent.com/tertiarycourses/datasets/master/iris.csv\"\n",
        "                     \n",
        "X = pd.read_csv(dataset_path)\n",
        "X"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "uvBNfoArZHJE",
        "colab_type": "text"
      },
      "source": [
        "### Split the Features and Target"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-WWug_KiSG1R",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "X = X.dropna()\n",
        "y = X.pop('species')\n",
        "y"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1g2xa1AfZLOc",
        "colab_type": "text"
      },
      "source": [
        "### Encode the Label"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-NHVS85_XwmI",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.preprocessing import LabelEncoder\n",
        "le = LabelEncoder()\n",
        "y = le.fit_transform(y)\n",
        "y"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "eWzBd3fbolrL",
        "colab_type": "text"
      },
      "source": [
        "### Scale/Normalize the Features"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7BHWhi9CnqXt",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Scale the input data\n",
        "from sklearn.preprocessing import MinMaxScaler\n",
        "\n",
        "scaler = MinMaxScaler()\n",
        "\n",
        "X_scaled = pd.DataFrame(scaler.fit_transform(X), columns=X.columns)\n",
        "X_scaled"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "g26os1ArZUP2",
        "colab_type": "text"
      },
      "source": [
        "### Split and Randomize Training and Testing Dataset"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pZmIM1dWSJoZ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Split and Randomize the data\n",
        "from sklearn.model_selection import train_test_split\n",
        "\n",
        "X_train,X_test,y_train,y_test = train_test_split(X_scaled,y,test_size=0.3,random_state=100)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ngjgrhKVSLR7",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "X_train"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CctTFPLPSMoY",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "y_train"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "RaN5jg-MSSZa",
        "colab_type": "text"
      },
      "source": [
        "## Step 2 Define the Model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "w7TCz6bzSP1J",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# KNN Classifier\n",
        "\n",
        "from sklearn.neighbors import KNeighborsClassifier\n",
        "clf = KNeighborsClassifier(n_neighbors=3)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "2vNTifjwSWyQ",
        "colab_type": "text"
      },
      "source": [
        "## Step 3 Train the Model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wIrcgqKMSUjx",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "clf.fit(X_train,y_train)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "mrXm1SeNSa5Z",
        "colab_type": "text"
      },
      "source": [
        "## Step 4 Evaluate the Model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qNM5eaPmSXNx",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "clf.score(X_test,y_test)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yVchIkxcHrtQ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "X_pred = [[5.1,3.5,1.4,0.2]]\n",
        "X_pred = scaler.transform(X_pred)\n",
        "clf.predict(X_pred)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xzWZVKCDSz7l",
        "colab_type": "text"
      },
      "source": [
        "## Step 5 Save the Model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ykGk7YM0Sy37",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.externals import joblib\n",
        "joblib.dump(clf, 'iris.pkl') "
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "yjs2GBDKcBMc",
        "colab_type": "text"
      },
      "source": [
        ""
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "So3J1JgDS5V3",
        "colab_type": "text"
      },
      "source": [
        "## Step 6 Load the Model for Inference"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "69j2g_0TS2rc",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.externals import joblib\n",
        "new_clf = joblib.load('iris.pkl')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xmd9Q8oQS64a",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import numpy as np\n",
        "\n",
        "X_new = np.array([[6.7,3.1,4.7,1.5]])\n",
        "y = new_clf.predict(scaler.transform(X_new))\n",
        "\n",
        "label = {0:'sentosa',1:'versicolor',2:'virginica'}\n",
        "print('The flower is ',label[y[0]])"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0kkkbueKSk8S",
        "colab_type": "text"
      },
      "source": [
        "## Ex: Classifiers"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "h_9K-GXYkEsi",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Logistic Regression Classifier\n",
        "\n",
        "from sklearn.linear_model import LogisticRegression \n",
        "clf = LogisticRegression()\n",
        "clf.fit(X_train,y_train)\n",
        "clf.score(X_test,y_test)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KF5O1LDiSit1",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# SVM Classifier\n",
        "\n",
        "from sklearn.svm import SVC \n",
        "clf = SVC()\n",
        "clf.fit(X_train,y_train)\n",
        "clf.score(X_test,y_test)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wP-vG7l4Sm4D",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Guassian Navie Bayes Classifer\n",
        "\n",
        "from sklearn.naive_bayes import GaussianNB\n",
        "clf = GaussianNB()\n",
        "clf.fit(X_train,y_train)\n",
        "clf.score(X_test,y_test)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ipHb9W8wSqUF",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Decision Tree Classifer\n",
        "\n",
        "from sklearn.tree import DecisionTreeClassifier\n",
        "clf = DecisionTreeClassifier()\n",
        "clf.fit(X_train,y_train)\n",
        "clf.score(X_test,y_test)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "poHC_FrSSoy9",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# SGD Classifer\n",
        "\n",
        "from sklearn.linear_model import SGDClassifier\n",
        "clf = SGDClassifier()\n",
        "clf.fit(X_train,y_train)\n",
        "clf.score(X_test,y_test)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ygbWV3AJSuhH",
        "colab_type": "text"
      },
      "source": [
        "## Ensemble Methods"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QX8dJhTGpA4R",
        "colab_type": "text"
      },
      "source": [
        "### Bagging"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0dspVnRMSry_",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Random Forest Classifier\n",
        "\n",
        "from sklearn.ensemble import RandomForestClassifier\n",
        "\n",
        "clf = RandomForestClassifier()\n",
        "clf.fit(X_train,y_train)\n",
        "clf.score(X_test,y_test)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "bqBhTJ8Huh46",
        "colab_type": "text"
      },
      "source": [
        "### Boosting"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nnggSRwoSxBQ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Gradient Boosting Classifier\n",
        "\n",
        "from sklearn.ensemble import GradientBoostingClassifier\n",
        "\n",
        "clf = GradientBoostingClassifier()\n",
        "clf.fit(X_train,y_train)\n",
        "clf.score(X_test,y_test)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "p3W8A0-LyR_D",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# AdaBoosting Classifier\n",
        "\n",
        "from sklearn.ensemble import AdaBoostClassifier\n",
        "\n",
        "clf = AdaBoostClassifier()\n",
        "clf.fit(X_train,y_train)\n",
        "clf.score(X_test,y_test)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "r43vOAxXwlZb",
        "colab_type": "text"
      },
      "source": [
        "### Stacking"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZF6SQwiwwFxl",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.linear_model import LogisticRegression\n",
        "from sklearn.naive_bayes import GaussianNB\n",
        "from sklearn.ensemble import RandomForestClassifier, VotingClassifier\n",
        "clf1 = LogisticRegression()\n",
        "clf2 = RandomForestClassifier(n_estimators=50, random_state=1)\n",
        "clf3 = GaussianNB()\n",
        "\n",
        "clf = VotingClassifier(estimators=[('lr', clf1), ('rf', clf2), ('gnb', clf3)], voting='hard')\n",
        "clf.fit(X_train,y_train)\n",
        "clf.score(X_test,y_test)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3rJToRg9S992",
        "colab_type": "text"
      },
      "source": [
        "## Ex: Emsemble Methods"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "oCazcLq9MGMx",
        "colab_type": "text"
      },
      "source": [
        "### Prepare the Data"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PI5-St7ZS8MO",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Load dataset\n",
        "import pandas as pd\n",
        "dataset_path = \"https://raw.githubusercontent.com/tertiarycourses/datasets/master/winequality-red.csv\"\n",
        "                     \n",
        "X = pd.read_csv(dataset_path,sep=';')\n",
        "X"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Ek8UfRBHex5U",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "X = X.dropna()\n",
        "y = X.pop('quality')\n",
        "y"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8QWasA3MozeF",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Scale the input data\n",
        "from sklearn.preprocessing import MinMaxScaler\n",
        "\n",
        "scaler = MinMaxScaler()\n",
        "\n",
        "X_scaled = pd.DataFrame(scaler.fit_transform(X), columns=X.columns)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lZRDf9b_hVKe",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Split and Randomize the data\n",
        "from sklearn.model_selection import train_test_split\n",
        "\n",
        "X_train,X_test,y_train,y_test = train_test_split(X,y,test_size=0.3,random_state=100)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "_Z2dP_lTMLno",
        "colab_type": "text"
      },
      "source": [
        "### Define, Train and Evalaute the Model\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2l44kAt0TGZf",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Decision Tree Classifer\n",
        "\n",
        "from sklearn import tree\n",
        "clf = tree.DecisionTreeClassifier()\n",
        "clf.fit(X_train,y_train)\n",
        "clf.score(X_test,y_test)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "llcDndTKTHjO",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Random Forest Classifier\n",
        "\n",
        "from sklearn.ensemble import RandomForestClassifier\n",
        "\n",
        "clf = RandomForestClassifier()\n",
        "clf.fit(X_train,y_train)\n",
        "clf.score(X_test,y_test)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Wxk4gAT5TJsh",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Gradient Boosting Tree Classifier\n",
        "\n",
        "from sklearn.ensemble import GradientBoostingClassifier\n",
        "\n",
        "clf = GradientBoostingClassifier()\n",
        "clf.fit(X_train,y_train)\n",
        "clf.score(X_test,y_test)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4VbBGS_GMQV2",
        "colab_type": "text"
      },
      "source": [
        ""
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HZmuvpbZyJC3",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Stacking Classifer\n",
        "from sklearn.linear_model import LogisticRegression\n",
        "from sklearn.naive_bayes import GaussianNB\n",
        "from sklearn.ensemble import RandomForestClassifier, VotingClassifier\n",
        "clf1 = LogisticRegression()\n",
        "clf2 = RandomForestClassifier(n_estimators=50, random_state=1)\n",
        "clf3 = GaussianNB()\n",
        "\n",
        "clf = VotingClassifier(estimators=[('lr', clf1), ('rf', clf2), ('gnb', clf3)], voting='hard')\n",
        "clf.fit(X_train,y_train)\n",
        "clf.score(X_test,y_test)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "IOKva-Jnh6et",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "i = 50\n",
        "x = scaler.transform(X_test.iloc[i:i+1])\n",
        "print('The predicted wine quality is ', clf.predict(x)[0])\n",
        "print('The actual wine quality is ',y_test.iloc[i])"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "YeIFDTHuTMkR",
        "colab_type": "text"
      },
      "source": [
        "## Confusion Matrix"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "J94RjKNpTLIr",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Load Iris dataset\n",
        "import pandas as pd\n",
        "dataset_path = \"https://raw.githubusercontent.com/tertiarycourses/datasets/master/iris.csv\"\n",
        "                     \n",
        "X = pd.read_csv(dataset_path)\n",
        "\n",
        "X = X.dropna()\n",
        "y = X.pop('species')\n",
        "\n",
        "from sklearn.preprocessing import LabelEncoder\n",
        "le = LabelEncoder()\n",
        "y = le.fit_transform(y)\n",
        "\n",
        "from sklearn.preprocessing import MinMaxScaler\n",
        "\n",
        "scaler = MinMaxScaler()\n",
        "X_scaled = pd.DataFrame(scaler.fit_transform(X), columns=X.columns)\n",
        "\n",
        "from sklearn.model_selection import train_test_split\n",
        "\n",
        "X_train,X_test,y_train,y_test = train_test_split(X_scaled,y,test_size=0.3,random_state=100)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5u6evXPRgOgY",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.neighbors import KNeighborsClassifier\n",
        "\n",
        "clf = KNeighborsClassifier(n_neighbors=3)\n",
        "clf.fit(X_test,y_test)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7CoK5V3YgS_p",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.metrics import confusion_matrix\n",
        "\n",
        "y_pred = clf.predict(X_test)\n",
        "print(confusion_matrix(y_test,y_pred))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jIKJ5dqmTRWA",
        "colab_type": "text"
      },
      "source": [
        "## Ex: Confusion Matrix"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_vCkcohpkzej",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Load Iris dataset\n",
        "import pandas as pd\n",
        "dataset_path = \"https://raw.githubusercontent.com/tertiarycourses/datasets/master/winequality-red.csv\"\n",
        "                     \n",
        "X = pd.read_csv(dataset_path,sep=';')\n",
        "\n",
        "X = X.dropna()\n",
        "y = X.pop('quality')\n",
        "\n",
        "from sklearn.preprocessing import LabelEncoder\n",
        "le = LabelEncoder()\n",
        "y = le.fit_transform(y)\n",
        "\n",
        "from sklearn.preprocessing import MinMaxScaler\n",
        "\n",
        "scaler = MinMaxScaler()\n",
        "X_scaled = pd.DataFrame(scaler.fit_transform(X), columns=X.columns)\n",
        "\n",
        "from sklearn.model_selection import train_test_split\n",
        "\n",
        "X_train,X_test,y_train,y_test = train_test_split(X_scaled,y,test_size=0.3,random_state=100)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HLB043QXTUY5",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# KNN Classifier\n",
        "\n",
        "from sklearn import neighbors\n",
        "clf = neighbors.KNeighborsClassifier()\n",
        "clf.fit(X_train,y_train)\n",
        "\n",
        "y_pred = clf.predict(X_test)\n",
        "confusion_matrix(y_test,y_pred)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7LNtjwpXlbcR",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Logistic Regression Classifier\n",
        "\n",
        "from sklearn.linear_model import LogisticRegression \n",
        "clf = LogisticRegression()\n",
        "clf.fit(X_train,y_train)\n",
        "\n",
        "y_pred = clf.predict(X_test)\n",
        "confusion_matrix(y_test,y_pred)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wJUCFqMsTNGQ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# SVM Classifier\n",
        "\n",
        "from sklearn import svm \n",
        "clf = svm.SVC()\n",
        "clf.fit(X_train,y_train)\n",
        "\n",
        "y_pred = clf.predict(X_test)\n",
        "confusion_matrix(y_test,y_pred)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2O2BUMnFlowv",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# GNB Classifier\n",
        "\n",
        "from sklearn.naive_bayes import GaussianNB\n",
        "clf = GaussianNB()\n",
        "clf.fit(X_train,y_train)\n",
        "\n",
        "y_pred = clf.predict(X_test)\n",
        "confusion_matrix(y_test,y_pred)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fUaD7KeAmLTy",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Decision Tree Classifer\n",
        "\n",
        "from sklearn.tree import DecisionTreeClassifier\n",
        "clf = DecisionTreeClassifier()\n",
        "clf.fit(X_train,y_train)\n",
        "\n",
        "y_pred = clf.predict(X_test)\n",
        "confusion_matrix(y_test,y_pred)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7W0xtvMqmQye",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# SGD Classifer\n",
        "\n",
        "from sklearn.linear_model import SGDClassifier\n",
        "clf = SGDClassifier()\n",
        "clf.fit(X_train,y_train)\n",
        "\n",
        "y_pred = clf.predict(X_test)\n",
        "confusion_matrix(y_test,y_pred)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QSLjmyCzvyEt",
        "colab_type": "text"
      },
      "source": [
        "## Metrics: Accuracy, Precision, Recall, F1 Scores"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8DpmGY-DxW_n",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Load Iris dataset\n",
        "import pandas as pd\n",
        "dataset_path = \"https://raw.githubusercontent.com/tertiarycourses/datasets/master/iris.csv\"\n",
        "                     \n",
        "X = pd.read_csv(dataset_path)\n",
        "\n",
        "X = X.dropna()\n",
        "y = X.pop('species')\n",
        "\n",
        "from sklearn.preprocessing import LabelEncoder\n",
        "le = LabelEncoder()\n",
        "y = le.fit_transform(y)\n",
        "y[y==2]=0\n",
        "\n",
        "from sklearn.preprocessing import MinMaxScaler\n",
        "\n",
        "scaler = MinMaxScaler()\n",
        "X_scaled = pd.DataFrame(scaler.fit_transform(X), columns=X.columns)\n",
        "\n",
        "from sklearn.model_selection import train_test_split\n",
        "\n",
        "X_train,X_test,y_train,y_test = train_test_split(X_scaled,y,test_size=0.3,random_state=100)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mMj0mD84TfNP",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# K Nearest Neighbor Classifier\n",
        "\n",
        "from sklearn.metrics import accuracy_score,precision_score,recall_score,f1_score\n",
        "\n",
        "from sklearn import neighbors\n",
        "clf = neighbors.KNeighborsClassifier()\n",
        "clf.fit(X_train,y_train)\n",
        "\n",
        "y_pred = clf.predict(X_test)\n",
        "print('Accuracy = ',accuracy_score(y_test,y_pred))\n",
        "print('Precision = ',precision_score(y_test,y_pred))\n",
        "print('Recall = ',recall_score(y_test,y_pred))\n",
        "print('F1 Score = ',f1_score(y_test,y_pred))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "j-1J30adTjRo",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Logistic Regression Classifier\n",
        "\n",
        "from sklearn.linear_model import LogisticRegression \n",
        "clf = LogisticRegression()\n",
        "clf.fit(X_train,y_train)\n",
        "\n",
        "y_pred = clf.predict(X_test)\n",
        "print('Accuracy = ',accuracy_score(y_test,y_pred))\n",
        "print('Precision = ',precision_score(y_test,y_pred))\n",
        "print('Recall = ',recall_score(y_test,y_pred))\n",
        "print('F1 Score = ',f1_score(y_test,y_pred))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qvdxJwfiyDG0",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# SVM Classifier\n",
        "\n",
        "from sklearn import svm \n",
        "clf = svm.SVC()\n",
        "clf.fit(X_train,y_train)\n",
        "\n",
        "y_pred = clf.predict(X_test)\n",
        "print('Accuracy = ',accuracy_score(y_test,y_pred))\n",
        "print('Precision = ',precision_score(y_test,y_pred))\n",
        "print('Recall = ',recall_score(y_test,y_pred))\n",
        "print('F1 Score = ',f1_score(y_test,y_pred))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "X-ABHzxDx6C5",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Decision Tree Classifer\n",
        "\n",
        "from sklearn.tree import DecisionTreeClassifier\n",
        "clf = DecisionTreeClassifier()\n",
        "clf.fit(X_train,y_train)\n",
        "\n",
        "y_pred = clf.predict(X_test)\n",
        "print('Accuracy = ',accuracy_score(y_test,y_pred))\n",
        "print('Precision = ',precision_score(y_test,y_pred))\n",
        "print('Recall = ',recall_score(y_test,y_pred))\n",
        "print('F1 Score = ',f1_score(y_test,y_pred))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "IbffgLnWTa3L",
        "colab_type": "text"
      },
      "source": [
        "### Classification Report"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "eJVqcq8qTXPO",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.metrics import classification_report\n",
        "print(classification_report(y_test, y_pred))\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "XzY_0yyKWMTm",
        "colab_type": "text"
      },
      "source": [
        "## Ex: : Metrics"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CNQTEbNOWLZI",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Load Wine Quality dataset\n",
        "\n",
        "import pandas as pd\n",
        "dataset_path = \"https://raw.githubusercontent.com/tertiarycourses/datasets/master/winequality-red.csv\"\n",
        "\n",
        "X = pd.read_csv(dataset_path,sep=\";\")\n",
        "\n",
        "X = X.dropna()\n",
        "y = X.pop('quality')\n",
        "\n",
        "y[y<6]=0\n",
        "y[y>=6]=1\n",
        "\n",
        "from sklearn.preprocessing import MinMaxScaler\n",
        "\n",
        "scaler = MinMaxScaler()\n",
        "X_scaled = pd.DataFrame(scaler.fit_transform(X), columns=X.columns)\n",
        "\n",
        "from sklearn.model_selection import train_test_split\n",
        "\n",
        "X_train,X_test,y_train,y_test = train_test_split(X_scaled,y,test_size=0.3,random_state=100)\n",
        "y_train"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7Oj0DQvjXmHx",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# KNN Classifier\n",
        "\n",
        "from sklearn.metrics import accuracy_score,precision_score,recall_score,f1_score\n",
        "from sklearn.metrics import classification_report\n",
        "\n",
        "from sklearn import neighbors\n",
        "clf = neighbors.KNeighborsClassifier()\n",
        "clf.fit(X_train,y_train)\n",
        "\n",
        "y_pred = clf.predict(X_test)\n",
        "print('Accuracy = ',accuracy_score(y_test,y_pred))\n",
        "print('Precision = ',precision_score(y_test,y_pred))\n",
        "print('Recall = ',recall_score(y_test,y_pred))\n",
        "print('F1 Score = ',f1_score(y_test,y_pred))\n",
        "\n",
        "print(classification_report(y_test, y_pred))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gVLHU-RvYAZ7",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# GNB Classifier\n",
        "\n",
        "clf = GaussianNB()\n",
        "clf.fit(X_train,y_train)\n",
        "clf.score(X_test,y_test)\n",
        "\n",
        "y_pred = clf.predict(X_test)\n",
        "print('Accuracy = ',accuracy_score(y_test,y_pred))\n",
        "print('Precision = ',precision_score(y_test,y_pred))\n",
        "print('Recall = ',recall_score(y_test,y_pred))\n",
        "print('F1 Score = ',f1_score(y_test,y_pred))\n",
        "\n",
        "print(classification_report(y_test, y_pred))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FLXW7zYUXuoV",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# SVM Classifier\n",
        "\n",
        "from sklearn import svm \n",
        "clf = svm.SVC()\n",
        "clf.fit(X_train,y_train)\n",
        "\n",
        "y_pred = clf.predict(X_test)\n",
        "print('Accuracy = ',accuracy_score(y_test,y_pred))\n",
        "print('Precision = ',precision_score(y_test,y_pred))\n",
        "print('Recall = ',recall_score(y_test,y_pred))\n",
        "print('F1 Score = ',f1_score(y_test,y_pred))\n",
        "\n",
        "print(classification_report(y_test, y_pred))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3ZFZ6AL_ztLC",
        "colab_type": "text"
      },
      "source": [
        "## ROC Curve and AUC"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CsiVdC3D9914",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Load Iris dataset\n",
        "import pandas as pd\n",
        "dataset_path = \"https://raw.githubusercontent.com/tertiarycourses/datasets/master/iris.csv\"\n",
        "                     \n",
        "X = pd.read_csv(dataset_path)\n",
        "\n",
        "X = X.dropna()\n",
        "y = X.pop('species')\n",
        "\n",
        "from sklearn.preprocessing import LabelEncoder\n",
        "le = LabelEncoder()\n",
        "y = le.fit_transform(y)\n",
        "y[y==2]=0\n",
        "\n",
        "from sklearn.preprocessing import MinMaxScaler\n",
        "\n",
        "scaler = MinMaxScaler()\n",
        "X_scaled = pd.DataFrame(scaler.fit_transform(X), columns=X.columns)\n",
        "\n",
        "from sklearn.model_selection import train_test_split\n",
        "\n",
        "X_train,X_test,y_train,y_test = train_test_split(X_scaled,y,test_size=0.3,random_state=100)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "MYCEWiM9Tk7Q",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.linear_model import LogisticRegression \n",
        "clf = LogisticRegression()\n",
        "\n",
        "clf.fit(X_train,y_train)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0bDoGu0Aj0QL",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.metrics import roc_curve,roc_auc_score\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "# predict probabilities\n",
        "y_probs = clf.predict_proba(X_test)\n",
        "\n",
        "# keep probabilities for the positive outcome only\n",
        "y_probs = y_probs[:, 1]\n",
        "\n",
        "fpr, tpr, threshold = roc_curve(y_test,y_probs)\n",
        "\n",
        "plt.plot(fpr,tpr,'o-')\n",
        "plt.xlabel('False Postive Rate')\n",
        "plt.ylabel('True Positive Rate')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4R4v9Ik3jUc1",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "auc_score = roc_auc_score(y_test, y_probs)\n",
        "print(\"AUC = \", auc_score)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Z3JKhbtq_bnw",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GYdNTRW-nu7j",
        "colab_type": "text"
      },
      "source": [
        "## Ex: ROC and AOC"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3temndcznuOy",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Load Wine Quality dataset\n",
        "\n",
        "import pandas as pd\n",
        "dataset_path = \"https://raw.githubusercontent.com/tertiarycourses/datasets/master/winequality-red.csv\"\n",
        "\n",
        "X = pd.read_csv(dataset_path,sep=\";\")\n",
        "\n",
        "X = X.dropna()\n",
        "y = X.pop('quality')\n",
        "\n",
        "y[y<6]=0\n",
        "y[y>=6]=1\n",
        "\n",
        "from sklearn.preprocessing import MinMaxScaler\n",
        "\n",
        "scaler = MinMaxScaler()\n",
        "X_scaled = pd.DataFrame(scaler.fit_transform(X), columns=X.columns)\n",
        "\n",
        "from sklearn.model_selection import train_test_split\n",
        "\n",
        "X_train,X_test,y_train,y_test = train_test_split(X_scaled,y,test_size=0.3,random_state=100)\n",
        "y_train"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HC2U9VEyjh5M",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.linear_model import LogisticRegression \n",
        "clf = LogisticRegression()\n",
        "\n",
        "clf.fit(X_train,y_train)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "epYlbSZzn0o2",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.metrics import roc_curve,roc_auc_score\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "# predict probabilities\n",
        "y_probs = clf.predict_proba(X_test)\n",
        "\n",
        "# keep probabilities for the positive outcome only\n",
        "y_probs = y_probs[:, 1]\n",
        "\n",
        "fpr, tpr, threshold = roc_curve(y_test,y_probs)\n",
        "\n",
        "plt.plot(fpr,tpr,'o-')\n",
        "plt.xlabel('False Postive Rate')\n",
        "plt.ylabel('True Positive Rate')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YJT2CFTMn4wT",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "auc_score = roc_auc_score(y_test, y_probs)\n",
        "print(\"AUC = \", auc_score)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QB9yRXS3T9oz",
        "colab_type": "text"
      },
      "source": [
        "# Topic 3 Regression"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "eRUtaba9gvSz",
        "colab_type": "text"
      },
      "source": [
        "## Load the Data"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "cLq7vbOBZH0e",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import pandas as pd\n",
        "dataset_path = \"https://raw.githubusercontent.com/selva86/datasets/master/BostonHousing.csv\"                     \n",
        "X = pd.read_csv(dataset_path)\n",
        "X"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "rO4a67Geg1o0",
        "colab_type": "text"
      },
      "source": [
        "### Prepare the Data"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DWdg6eXCjKiS",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Remove missing data\n",
        "X = X.dropna()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fbbJf-6xZnOW",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Define input and output\n",
        "y = X.pop('medv')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CUiXEwM8aI4V",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.preprocessing import MinMaxScaler\n",
        "\n",
        "scaler = MinMaxScaler()\n",
        "X_scaled = pd.DataFrame(scaler.fit_transform(X), columns=X.columns)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "g9_W8i5gZ0Rl",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Split and Randomize the data\n",
        "from sklearn.model_selection import train_test_split\n",
        "\n",
        "X_train,X_test,y_train,y_test = train_test_split(X_scaled,y,test_size=0.3,random_state=100)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "oj_wIAH-aeVA",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "X_train"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "tXOjGqU9hFws",
        "colab_type": "text"
      },
      "source": [
        "## Define the Model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_oSbX8TJa1fs",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Linear Regression Model\n",
        "\n",
        "from sklearn.linear_model import LinearRegression \n",
        "lm = LinearRegression()\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Sim2PEBohIzn",
        "colab_type": "text"
      },
      "source": [
        "### Train the Model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wbNsbDfDhPGp",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "lm.fit(X_train,y_train)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "n3_kH4VXhQ8i",
        "colab_type": "text"
      },
      "source": [
        "## Evaluate the Model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DBte2QcixjT3",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.metrics import mean_squared_error, r2_score\n",
        "\n",
        "yhat = lm.predict(X_test)\n",
        "mse = mean_squared_error(y_test,yhat)\n",
        "print('Mean Squared Error, Testing: ',mse)\n",
        "rsq = r2_score(y_test,yhat)\n",
        "print('R-square, Testing: ',rsq)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6zsyRMNUcK4i",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "%matplotlib inline\n",
        "import matplotlib.pyplot as plt \n",
        "\n",
        "yhat = lm.predict(X_test)\n",
        "\n",
        "plt.scatter(y_test,yhat)\n",
        "plt.xlabel('Actual Housing Price')\n",
        "plt.ylabel('Predicted Price')\n",
        "plt.plot([0, 50], [0, 50],'r')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "suUk6fWwUEdD",
        "colab_type": "text"
      },
      "source": [
        "## Ex: Linear Regression"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fn9NdXzhUBhL",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import pandas as pd\n",
        "dataset_path = \"https://raw.githubusercontent.com/tertiarycourses/datasets/master/iris.csv\"\n",
        "                     \n",
        "X = pd.read_csv(dataset_path)\n",
        "X"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "00O99i_vUGKF",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "X = X.dropna()\n",
        "X.pop('species')\n",
        "X"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "q998RnZFe-4P",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "y = X.pop('sepal_width')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KXpmECprfHwb",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.preprocessing import MinMaxScaler\n",
        "\n",
        "scaler = MinMaxScaler()\n",
        "X_scaled = pd.DataFrame(scaler.fit_transform(X), columns=X.columns)\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NSTtK0LMfC6Z",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Split and Randomize the data\n",
        "from sklearn.model_selection import train_test_split\n",
        "\n",
        "X_train,X_test,y_train,y_test = train_test_split(X_scaled,y,test_size=0.3,random_state=100)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jDAB6i3GfQTU",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Linear Regression Model\n",
        "\n",
        "from sklearn.linear_model import LinearRegression\n",
        "\n",
        "lm = LinearRegression()\n",
        "lm.fit(X_train,y_train)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DdAXEFsoz8pT",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.metrics import mean_squared_error, r2_score\n",
        "\n",
        "yhat = lm.predict(X_test)\n",
        "mse = mean_squared_error(y_test,yhat)\n",
        "print('Mean Squared Error, Testing: ',mse)\n",
        "rsq = r2_score(y_test,yhat)\n",
        "print('R-square, Testing: ',rsq)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dDS_s-GMUH21",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import matplotlib.pyplot as plt \n",
        "\n",
        "yhat = lm.predict(X_test)\n",
        "\n",
        "plt.scatter(y_test,yhat)\n",
        "plt.xlabel('Actual Sepal Width')\n",
        "plt.ylabel('Predicted Sepal Width')\n",
        "plt.plot([0, 5], [0, 5],'r')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Kxl_mlFj7TZQ",
        "colab_type": "text"
      },
      "source": [
        "## Regularizations"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "bKrJov9XWM84",
        "colab_type": "text"
      },
      "source": [
        "### Ridge Regularization"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8mV4HQFA3-wp",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.linear_model import Ridge\n",
        "rr = Ridge(alpha=0.1) \n",
        "rr.fit(X_train,y_train)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mNJZy6Fs4TJW",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.metrics import mean_squared_error, r2_score\n",
        "\n",
        "yhat = rr.predict(X_\n",
        "                  test)\n",
        "mse = mean_squared_error(y_test,yhat)\n",
        "print('Mean Squared Error, Testing: ',mse)\n",
        "rsq = r2_score(y_test,yhat)\n",
        "print('R-square, Testing: ',rsq)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ABp5KcZAWQsg",
        "colab_type": "text"
      },
      "source": [
        "### Lasso Regularizaton"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hE6H9xG54b01",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.linear_model import Lasso\n",
        "lr = Lasso(alpha=0.01) \n",
        "lr.fit(X_train,y_train)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6d5v4nnd4hxc",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.metrics import mean_squared_error, r2_score\n",
        "\n",
        "yhat = lr.predict(X_test)\n",
        "mse = mean_squared_error(y_test,yhat)\n",
        "print('Mean Squared Error, Testing: ',mse)\n",
        "rsq = r2_score(y_test,yhat)\n",
        "print('R-square, Testing: ',rsq)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "mIF29vAzWUAs",
        "colab_type": "text"
      },
      "source": [
        "### Elastic Net Regularization"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fuJHKNOo52sT",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.linear_model import ElasticNet\n",
        "er = ElasticNet(alpha=0.01,l1_ratio=0.5)\n",
        "er.fit(X_train,y_train)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XArePwP96NuV",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.metrics import mean_squared_error, r2_score\n",
        "\n",
        "yhat = er.predict(X_test)\n",
        "mse = mean_squared_error(y_test,yhat)\n",
        "print('Mean Squared Error, Testing: ',mse)\n",
        "rsq = r2_score(y_test,yhat)\n",
        "print('R-square, Testing: ',rsq)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9L6qfnFmUokI",
        "colab_type": "text"
      },
      "source": [
        "# Topic 4 Clustering"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gV4lHfgpUtFb",
        "colab_type": "text"
      },
      "source": [
        "## Load the Data"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CIurmsPdUuwp",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.datasets.samples_generator import make_blobs\n",
        "\n",
        "centers = [[1,1],[1.5,1.5],[2,2]]\n",
        "X,y = make_blobs(n_samples=100,centers = centers,cluster_std=0.1)\n",
        "\n",
        "import matplotlib.pyplot as plt\n",
        "plt.scatter(X[:,0],X[:,1])"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "TvjMG0u9UyND",
        "colab_type": "text"
      },
      "source": [
        "## load the Model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NQnnWHsdUwG7",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.cluster import KMeans\n",
        "\n",
        "cluster = KMeans(n_clusters=3,random_state=10)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "oGmfEOM1U29E",
        "colab_type": "text"
      },
      "source": [
        "## Train the Model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "53woiFbEUytx",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "cluster.fit(X)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ozO-cSZxU9mn",
        "colab_type": "text"
      },
      "source": [
        "## Evaluate the Model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "cNLbF72kU6dz",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "cluster.labels_"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gzwKnYuiU_g9",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "cluster.cluster_centers_"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ImTBYFgcVBE_",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "plt.figure(figsize=(10,5))\n",
        "plt.subplot(1,2,1)\n",
        "plt.scatter(X[:,0],X[:,1],c=y)\n",
        "plt.title(\"Actual\")\n",
        "plt.subplot(1,2,2)\n",
        "plt.scatter(X[:,0],X[:,1],c=cluster.labels_)\n",
        "plt.title(\"Prediction\")"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DBoLBR4gwVJw",
        "colab_type": "text"
      },
      "source": [
        "## Silhouette Analysis"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NXOP3fYIhJ1-",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.metrics import silhouette_score\n",
        "\n",
        "for i in range(2,10):\n",
        "    cluster = KMeans(n_clusters=i,random_state=10)\n",
        "    cluster.fit(X)\n",
        "    s = silhouette_score(X, cluster.labels_)\n",
        "    print('Cluster: ',i, 'Silhouette Score :',s )"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "g6WqUDJGVIJl",
        "colab_type": "text"
      },
      "source": [
        "## Ex: KMeans Clustering"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "cOtoqQZLF85O",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Load Iris dataset\n",
        "import pandas as pd\n",
        "dataset_path = \"https://raw.githubusercontent.com/tertiarycourses/datasets/master/iris.csv\"\n",
        "                     \n",
        "X = pd.read_csv(dataset_path)\n",
        "\n",
        "X = X.dropna()\n",
        "y = X.pop('species')\n",
        "\n",
        "from sklearn.preprocessing import LabelEncoder\n",
        "le = LabelEncoder()\n",
        "y = le.fit_transform(y)\n",
        "X"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GZQo7BIvkeML",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.preprocessing import StandardScaler\n",
        "\n",
        "scaler = StandardScaler()\n",
        "X_scaled = pd.DataFrame(scaler.fit_transform(X), columns=X.columns)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VuQO7MXJwlFh",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.cluster import KMeans\n",
        "from sklearn.metrics import silhouette_score\n",
        "\n",
        "for i in range(2,10):\n",
        "    cluster = KMeans(n_clusters=i,random_state=10)\n",
        "    cluster.fit(X_scaled)\n",
        "    s = silhouette_score(X_scaled, cluster.labels_)\n",
        "    print('Cluster: ',i, 'Silhouette Score :',s )"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rGiGn-TqVC17",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.cluster import KMeans\n",
        "\n",
        "cluster = KMeans(n_clusters=3,random_state=10)\n",
        "cluster.fit(X)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NPXlaUkYNzgA",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import matplotlib.pyplot as plt \n",
        "\n",
        "plt.figure(figsize=(10,5))\n",
        "x1 = 'petal_length'\n",
        "x2 = 'petal_width'\n",
        "plt.subplot(1,2,1)\n",
        "plt.scatter(X[x1],X[x2],c=y)\n",
        "plt.title(\"Actual\")\n",
        "plt.subplot(1,2,2)\n",
        "plt.scatter(X[x1],X[x2],c=cluster.labels_)\n",
        "plt.title(\"Prediction\")"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jLhWtO8bVR7G",
        "colab_type": "text"
      },
      "source": [
        "## Hierarchical Agglormerative Clustering"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7cDJsbaBI_JV",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.datasets.samples_generator import make_blobs\n",
        "\n",
        "centers = [[1,1],[1.5,1.5],[2,2]]\n",
        "X,y = make_blobs(n_samples=100,centers = centers,cluster_std=0.1)\n",
        "\n",
        "import matplotlib.pyplot as plt\n",
        "plt.scatter(X[:,0],X[:,1])"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yOWsnsTbLlB5",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from scipy.cluster.hierarchy import dendrogram, linkage\n",
        "\n",
        "Z = linkage(X)\n",
        "d = dendrogram(Z)\n",
        "plt.plot(d)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nK8sj8y4JCtX",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.cluster import AgglomerativeClustering\n",
        "\n",
        "cluster = AgglomerativeClustering(n_clusters = 3)\n",
        "cluster.fit(X)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "P9mELgyyJJ-9",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "plt.figure(figsize=(20,10))\n",
        "plt.subplot(1,2,1)\n",
        "plt.scatter(X[:,0],X[:,1],c=y)\n",
        "plt.title(\"Actual\")\n",
        "plt.subplot(1,2,2)\n",
        "plt.scatter(X[:,0],X[:,1],c=cluster.labels_)\n",
        "plt.title(\"Prediction\")"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "mE5GrTSsRGOQ",
        "colab_type": "text"
      },
      "source": [
        "## Ex: Hierarchical Agglormerative Clustering"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XtVNBx4AM8ZZ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Load Iris dataset\n",
        "import pandas as pd\n",
        "dataset_path = \"https://raw.githubusercontent.com/tertiarycourses/datasets/master/iris.csv\"\n",
        "                     \n",
        "X = pd.read_csv(dataset_path)\n",
        "\n",
        "X = X.dropna()\n",
        "y = X.pop('species')\n",
        "\n",
        "from sklearn.preprocessing import LabelEncoder\n",
        "le = LabelEncoder()\n",
        "y = le.fit_transform(y)\n",
        "X"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "TH1GmAVIw23p",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.preprocessing import StandardScaler\n",
        "\n",
        "scaler = StandardScaler()\n",
        "X_scaled = pd.DataFrame(scaler.fit_transform(X), columns=X.columns)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ibzuRq-MNRf9",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from scipy.cluster.hierarchy import dendrogram, linkage\n",
        "\n",
        "Z = linkage(X_scaled)\n",
        "d = dendrogram(Z)\n",
        "plt.plot(d)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SiaNRwkDNl1G",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.cluster import AgglomerativeClustering\n",
        "\n",
        "cluster = AgglomerativeClustering(n_clusters = 3)\n",
        "cluster.fit(X_scaled)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "giELkzehNoN_",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import matplotlib.pyplot as plt \n",
        "\n",
        "plt.figure(figsize=(20,10))\n",
        "x1 = 'petal_length'\n",
        "x2 = 'petal_width'\n",
        "plt.subplot(1,2,1)\n",
        "plt.scatter(X[x1],X[x2],c=y)\n",
        "plt.title(\"Actual\")\n",
        "plt.subplot(1,2,2)\n",
        "plt.scatter(X[x1],X[x2],c=cluster.labels_)\n",
        "plt.title(\"Prediction\")"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gFzO6Xj-Vdrc",
        "colab_type": "text"
      },
      "source": [
        "# Topic 5 Dimension Reduction"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fZg4O88efglN",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import numpy as np\n",
        "\n",
        "X = np.array([[-1, -1], [-2, -1], [-3, -2], [1, 1], [2, 1], [3, 2]])\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "z4K_7HU9T9pl",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "\n",
        "plt.scatter(X[:,0],X[:,1])\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3WVFQBI4UBfQ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn import decomposition\n",
        "pca = decomposition.PCA(n_components=2)\n",
        "pca.fit_transform(X)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QddKazgnflsD",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "pca.explained_variance_"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QVABVf6Yfp_B",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "pca.explained_variance_ratio_"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DloKXymeUEYo",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "pca.components_"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "l20Y8JccfsbP",
        "colab_type": "text"
      },
      "source": [
        "## PCA on Iris Dataset"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "tzHPt80TWqDT",
        "colab_type": "text"
      },
      "source": [
        "## Exercise: PCA on K-Means Clustering"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pkPLY5rsVS5n",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 419
        },
        "outputId": "dd19b012-3451-4a74-9734-6e89cca32c87"
      },
      "source": [
        "# Load Iris dataset\n",
        "import pandas as pd\n",
        "dataset_path = \"https://raw.githubusercontent.com/tertiarycourses/datasets/master/iris.csv\"\n",
        "                     \n",
        "X = pd.read_csv(dataset_path)\n",
        "\n",
        "X = X.dropna()\n",
        "y = X.pop('species')\n",
        "\n",
        "from sklearn.preprocessing import LabelEncoder\n",
        "le = LabelEncoder()\n",
        "y = le.fit_transform(y)\n",
        "X"
      ],
      "execution_count": 40,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<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>sepal_length</th>\n",
              "      <th>sepal_width</th>\n",
              "      <th>petal_length</th>\n",
              "      <th>petal_width</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>5.1</td>\n",
              "      <td>3.5</td>\n",
              "      <td>1.4</td>\n",
              "      <td>0.2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>4.9</td>\n",
              "      <td>3.0</td>\n",
              "      <td>1.4</td>\n",
              "      <td>0.2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>4.7</td>\n",
              "      <td>3.2</td>\n",
              "      <td>1.3</td>\n",
              "      <td>0.2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>4.6</td>\n",
              "      <td>3.1</td>\n",
              "      <td>1.5</td>\n",
              "      <td>0.2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>5.0</td>\n",
              "      <td>3.6</td>\n",
              "      <td>1.4</td>\n",
              "      <td>0.2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>145</th>\n",
              "      <td>6.7</td>\n",
              "      <td>3.0</td>\n",
              "      <td>5.2</td>\n",
              "      <td>2.3</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>146</th>\n",
              "      <td>6.3</td>\n",
              "      <td>2.5</td>\n",
              "      <td>5.0</td>\n",
              "      <td>1.9</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>147</th>\n",
              "      <td>6.5</td>\n",
              "      <td>3.0</td>\n",
              "      <td>5.2</td>\n",
              "      <td>2.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>148</th>\n",
              "      <td>6.2</td>\n",
              "      <td>3.4</td>\n",
              "      <td>5.4</td>\n",
              "      <td>2.3</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>149</th>\n",
              "      <td>5.9</td>\n",
              "      <td>3.0</td>\n",
              "      <td>5.1</td>\n",
              "      <td>1.8</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>150 rows × 4 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "     sepal_length  sepal_width  petal_length  petal_width\n",
              "0             5.1          3.5           1.4          0.2\n",
              "1             4.9          3.0           1.4          0.2\n",
              "2             4.7          3.2           1.3          0.2\n",
              "3             4.6          3.1           1.5          0.2\n",
              "4             5.0          3.6           1.4          0.2\n",
              "..            ...          ...           ...          ...\n",
              "145           6.7          3.0           5.2          2.3\n",
              "146           6.3          2.5           5.0          1.9\n",
              "147           6.5          3.0           5.2          2.0\n",
              "148           6.2          3.4           5.4          2.3\n",
              "149           5.9          3.0           5.1          1.8\n",
              "\n",
              "[150 rows x 4 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 40
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7QFmOAoCx1m2",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.preprocessing import StandardScaler\n",
        "\n",
        "scaler = StandardScaler()\n",
        "X_scaled = pd.DataFrame(scaler.fit_transform(X), columns=X.columns)"
      ],
      "execution_count": 41,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_6G9K-ODWiXx",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.decomposition import PCA\n",
        "pca = PCA(n_components=2)"
      ],
      "execution_count": 42,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "52UC7ZibWjti",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "X_t= pca.fit_transform(X_scaled)"
      ],
      "execution_count": 43,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "H_l3PEXJWlhb",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "outputId": "582ad91a-34ff-4dd9-ca6d-b5ed164e4b09"
      },
      "source": [
        "pca.explained_variance_"
      ],
      "execution_count": 44,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([2.93035378, 0.92740362])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 44
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dRbKq_bIWmw_",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "outputId": "989a8ec0-9a48-4f8e-c3fc-cec6d610b295"
      },
      "source": [
        "pca.explained_variance_ratio_"
      ],
      "execution_count": 45,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([0.72770452, 0.23030523])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 45
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1DSD10mjWoLJ",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 71
        },
        "outputId": "383a58ad-9d00-49de-c91f-2437af8cbfa6"
      },
      "source": [
        "from sklearn.cluster import KMeans\n",
        "\n",
        "cluster = KMeans(n_clusters=3)\n",
        "cluster.fit(X_scaled)\n",
        "\n",
        "cluster2 = KMeans(n_clusters=3)\n",
        "cluster2.fit(X_t)"
      ],
      "execution_count": 48,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "KMeans(algorithm='auto', copy_x=True, init='k-means++', max_iter=300,\n",
              "       n_clusters=3, n_init=10, n_jobs=None, precompute_distances='auto',\n",
              "       random_state=None, tol=0.0001, verbose=0)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 48
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xrVqFfjbWtGR",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 911
        },
        "outputId": "0d4d13ac-97b8-464f-eaf4-9a7b4b2a8bff"
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "\n",
        "x1 = 'petal_length'\n",
        "x2 = 'petal_width'\n",
        "\n",
        "plt.figure(figsize=(15,15))\n",
        "plt.subplot(2,2,1)\n",
        "plt.scatter(X[x1],X[x2],c=y)\n",
        "plt.title(\"Actual\")\n",
        "plt.xlabel(x1)\n",
        "plt.ylabel(x2)\n",
        "plt.subplot(2,2,2)\n",
        "plt.scatter(X[x1],X[x2],c=cluster.labels_)\n",
        "plt.title(\"Actual Clustering Prediction\")\n",
        "plt.xlabel(x1)\n",
        "plt.ylabel(x2)\n",
        "\n",
        "plt.subplot(2,2,3)\n",
        "plt.scatter(X_t[:,0],X_t[:,1],c=y)\n",
        "plt.title('PCA')\n",
        "plt.xlabel('P1')\n",
        "plt.ylabel('P2')\n",
        "\n",
        "plt.subplot(2,2,4)\n",
        "plt.scatter(X_t[:,0],X_t[:,1],c=cluster2.labels_)\n",
        "plt.title('PCA Clustering Prediction')\n",
        "plt.xlabel('P1')\n",
        "plt.ylabel('P2')"
      ],
      "execution_count": 53,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0, 0.5, 'P2')"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 53
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1080x1080 with 4 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "w3u4iYWkWufS",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}