Performance Analysis and Prediction in Grassroots Football: The Use of GPS Analytics, Machine Learning, and Deep Learning

This study examines the application of machine learning and deep learning techniques for performance monitoring and prediction in grassroots football. Using GPS tracking data collected over an entire season, we analyze player movements, heatmaps and high-speed running activities during training and competitive matches. The research focuses on two playing positions: Central Midfielder and Left Wing. We implement six machine learning models to predict player performance and compare their accuracies. Our findings reveal significant differences in physical demands between match and training sessions across playing positions. The study demonstrates the potential of data analytics in informing player development, detecting injury risks, and enhancing decision-making in grassroots football.