XGBoost
XGBoost performed best among the tested regression methods and was selected for the final models. The final models clipped predictions outside the valid ODI and NRS ranges. The study selected 22 predictors after SHAP analysis and expert re…
1 sources - 4 claims
XGBoost performed best among the tested regression methods and was selected for the final models. The final models clipped predictions outside the valid ODI and NRS ranges. The study selected 22 predictors after SHAP analysis and expert review for clinical relevance and legality. XGBoost was described as suitable for structured data but requiring hyperparameter tuning to reduce overfitting risk.