Football Match Prediction Project Using Random Forest (RF) Classifier

Table of Contents

1. Project Overview
2. Repository Contents * Web Scraping * Model Development
3. Model Evaluation
4. Results
5. Future Improvements

Project Overview

In this project, I harnessed machine learning classifier called Random Forest classifier to predict football match outcomes. This choice was driven by the classifier's strength in capturing non-linear relationships within the data, an important aspect in predicting football match results. This project aims to provide insights into the world of football match forecasting. Whether you are a sports enthusiast, a sports bettor, or simply curious about the power of data and machine learning in predicting football match results, this project provides valuable insights.

Repository Contents

Web Scraping
web scraping premier league data.ipynb: This file scrapes historical football match data from fbref, cleans it meticulously, and saves it as a CSV file for further analysis and model development.

Model Development
Matches_prediction_model.ipynb: The main script. It imports the pre-processed match data, conducts data wrangling, analysis, and feature extraction. Features are engineered to capture essential factors affecting match outcomes. The Random Forest classifier is used to build a predictive model, and GridSearchCV fine-tunes its hyperparameters for optimal performance.

Model Evaluation

Our final Random Forest classifier boasts a precision score of 60%. Which means the model was correct 60% of the time. Precision measures the accuracy of positive predictions in football match outcomes, meaning our model correctly predicts winning teams with a 60% accuracy rate.

Results

Our Random Forest classifier has demonstrated a precision score of 60%, showcasing its capability to predict football match outcomes accurately. This project offers valuable insights into the intersection of sports analytics and machine learning, highlighting the power of data-driven predictions in the world of football. Enjoy exploring and predicting the beautiful game!

Future Improvements

There are many areas this project can be improved on to provide better predictions for upcoming fixtures:

1. Aligning for deployment
2. Considering relative team strength
3. Ranking teams in order of strength