Table 2.

Mean performance metrics across all the cross-validation runs for each model with confidence intervals (CIs)

Model	ROC AUC (95% CI)	AUPRC (95% CI)	Brier score loss (95% CI)
LightGBM	0.8349 (0.8272, 0.8427)	0.1136 (0.1057, 0.1216)	0.0834 (0.0806, 0.0862)
XGBoost	0.8178 (0.8114, 0.8242)	0.1077 (0.0953, 0.1200)	0.0838 (0.0811, 0.0864)
Random forest	0.8363 (0.8236, 0.8491)	0.1182 (0.1079, 0.1285)	0.0915 (0.0855, 0.0975)
Logistic regression	0.8357 (0.8283, 0.8430)	0.1037 (0.0953, 0.1122)	0.2943 (0.2896, 0.2990)
Neural Network	0.8222 (0.8130, 0.8314)	0.1028 (0.0874, 0.1181)	0.2663 (0.1650, 0.3676)

ROC, receiver operating characteristic; AUC, area under the curve; AUPRC, area under the precision-recall curve.