Моделювання ризику травмування футболістів у короткостроковій перспективі на основі спортивних навантажень

Abstract

This research applies machine learning to predict short-term injury risks in professional football, addressing the need for proactive prevention tools. By integrating game events and injury histories, the study estimates non-contact injury probabilities within a 14-day window. Preliminary results show that Support Vector Machines (SVM) effectively identify high-risk scenarios using intensity metrics like pressure and duels. Furthermore, gender-specific modeling and the potential use of neural networks are discussed to further enhance predictive accuracy and capture complex performance patterns in player data