Using Machine Learning to Predict Premier League Matches

This project is a fullstack machine learning application that predicts Premier League football match outcomes using a Random Forest classifier trained on 4,788 historical matches from 2019-2024. The project uses engineered feature metrics and recent form with a React frontend and Flask backend to provide real-time predictions for current matchdays and custom matchups. It achieves 56.58% accuracy which is significantly better than random guessing (33.3%) and comparable to professional sports analysts (55-60%).