Fig. 5

External validation of Random Forest classifiers on GSE186334. (A) Fold-specific accuracy, recall, precision, F1, and ROC-AUC values are summarized as boxplots. (B) Precisionārecall curves for external validation folds and ensemble model. (C) ROC curves showing fold-specific AUCs (0.89ā0.94) and ensemble AUC (0.93). (D) Comparison of internal versus external validation metrics, highlighting modest performance reductions in external testing.