Fig. 3From: A machine learning-based prediction model for postoperative delirium in cardiac valve surgery using electronic health recordsComparison of AUROCs among machine learning models under full feature set (A) and selected feature set (C). Comparison of PR-Curves among machine learning models under full feature set (B) and selected feature set (D). Random Forest Classifier yielded the greatest AUC and AP in both feature setBack to article page