Supervised Machine Learning Methods in Predicting English Premier League Game Outcomes

This paper explores the application of machine learning (ML) models to predict soccer game outcomes in the English Premier League. Utilizing data from fbref.com, various Features and target variables were analyzed to assess their influence on game results. The study implemented multiple ML algorithms, including logistic regression, Naive Bayes, decision trees, k-nearest neighbors, random forest, AdaBoost, and multilayer perceptron. The results highlighted significant variations in prediction accuracy across different teams and models, with the Random Forest model achieving the highest average accuracy. The findings underscore the importance of careful algorithm selection and data processing to enhance prediction precision. This research contributes to the field of sports analytics, providing insights that can be applied to improve tactical planning and decision-making in sports. Future work will focus on optimizing models and exploring additional data sources to further increase prediction accuracy.