Translational Evaluation of Interpretable Machine Learning for Cardiovascular Risk Prediction: Calibration Decomposition, Subgroup Audits, and Decision-Utility Analysis
An evaluation of 308,774 records compared cardiovascular-risk models. Histogram gradient boosting achieved PR-AUC 0.3177, AUROC 0.8407 and Brier score 0.0633, alongside subgroup calibration and decision-utility audits.