ACL Injury Risk Predictor
Published:
A machine learning pipeline for predicting ACL injury risk in athletes from biomechanical measurements and performance metrics. Combines sports science domain knowledge with gradient boosting models and SHAP-based interpretability to produce actionable risk scores for clinicians and coaches.
Key work: Feature engineering from biomechanical measurements (force plate data, motion capture) · Gradient boosting classifier with calibrated probability outputs · SHAP values for per-athlete risk factor attribution · Cross-validation on longitudinal athlete datasets · Calibration analysis to ensure meaningful risk score interpretation
Stack: Python, XGBoost, scikit-learn, SHAP, sports analytics, biomechanics