AI-Assisted Prediction Models for Primary Anterior Cruciate Ligament Injury Risk and Recovery Time Using Basketball Athlete Performance Data

ACL injuries are among the most prevalent and debilitating injuries. This research aims to improve ACL injury risk stratification and recovery time prediction through AI, identify key risk factors, and provide a reproducible framework for clinical and sports settings. Methods involved data collection, Python plotting, and machine learning. Basketball athlete data from two Kaggle ACL injury datasets were cleaned and organized. Pivot tables and plots, including line, bar, and pie charts, were created in Google Sheets and refined in Python. Machine learning models, including Random Forest and Gradient Boosting, were trained and tested for predicting injury risk and recovery time. Random Forest and Logistic Regression calculated accuracy, precision, recall, and ROC-AUC for injury risk; Random Forest, Gradient Boosting, and Linear Regression calculated RMSE, MAE, and R² for recovery time. Factors increasing injury risk were height, weight, training intensity, and hours; factors decreasing risk were age and rest days. Longer recovery times were linked to age, height, and weight; jump height, speed, and rehabilitation efficiency reduced recovery time. Female basketball players had a higher injury risk. Flexibility exercises were least effective, while physiotherapy was most effective. The optimized Random Forest Classifier achieved perfect injury risk classification, outperforming Logistic Regression and Gradient Boosting. These results highlight the need for targeted prevention exercises, more physiotherapy programs to improve jump height and speed, and the use of Random Forest for optimal predictions. Future studies should explore prevention exercises, physiotherapy benefits, and Random Forest efficacy.