Machine learning-based prediction model for sports injury risk in biomechanics: A case study of joint injuries in basketball at a university in Xi’an

Basketball players are prone to joint injuries due to the sport’s high intensity and physical demands. Early prediction of injury risk is crucial for implementing effective prevention strategies. Incorporating biomechanics, this study focuses on basketball players at a university in Xi’an, China, aiming to develop a machine learning-based model to predict joint injury risk using easily collectable data such as training load, fatigue levels, and previous injury history. Considering regional differences, we observed that local and northern Chinese students are generally taller, while students from southern China are typically shorter. This anthropometric variation was included in our sampling and analysis. Utilizing data from 100 basketball players, the Random Forest algorithm achieved the best predictive performance with an accuracy of 85%. Key risk factors identified include high training load, elevated subjective fatigue scores, and a history of previous joint injuries. Additionally, biomechanical data were integrated to elucidate the underlying mechanisms of joint injuries, and the cellular responses to injury were explored. The results demonstrate that even with limited data types, machine learning methods can effectively predict joint injury risk among basketball players, providing a valuable tool for injury prevention.