Classification of Injury Risk Among Basketball Players: A Statistical Perspective

Injuries are a major concern in basketball, often caused by fatigue, high training loads, and insufficient recovery. This study developed a predictive model using player data-including demographics, training intensity, recovery patterns, and fatigue levels-to assess injury risk. Among several methods tested, binary logistic regression performed best, achieving 65% sensitivity, 78% specificity, and an AUC of 0.726. Key predictors included fatigue score, training hours, and recovery days. Fatigue emerged as the strongest risk factor, increasing injury odds by 3.52 times per unit rise, while each additional recovery day reduced the risk by 92%. Anthropometric variables like age, height, and weight showed no significant influence. Linear Discriminant Analysis (LDA) further confirmed moderate separation between injured and non-injured players. These findings highlight the importance of managing fatigue and recovery to reduce injury rates in competitive basketball. Future research can improve model accuracy through real-time monitoring and integration of advanced machine learning techniques.