Longitudinal study of match load zones and performance in elite football: a multivariate analysis and machine learning technique

Introduction: Recent technological advancements have revolutionized athlete monitoring, with football clubs worldwide leveraging embedded sensors to track player performance. However, interpreting the vast and complex data from these sensors remains a challenge for coaches and analysts. Objective: This study aims to identify the most significant match loads influencing player performance. Methodology: Data were collected from Terengganu Football Club (TFC) during the 2022 Malaysia Super League season. The Louvain clustering algorithm was employed to classify player performance levels, while logistic regression (logit model) identified key load zones linked to different performance tiers. The Kruskal-Wallis test was used to validate distinctions among these clusters. Results: Data were collected from Terengganu Football Club (TFC) during the 2022 Malaysia Super League season. The Louvain clustering algorithm was employed to classify player performance levels, while logistic regression (logit model) identified key load zones linked to different performance tiers. The Kruskal-Wallis test was used to validate distinctions among these clusters. Discussion: The findings provide insights into key load-related metrics, helping coaches understand their impact on player performance. These insights can guide workload management and training adjustments to enhance player efficiency and reduce injury risks. Conclusions: This study offers valuable guidance for refining training programs and developing more effective strategies to optimize team performance in modern football.