Table 3.

LR and XGBoost models performance parameters.

Models	AUC	Sensitivity	Specificity	Accuracy	Precision	Recall	F1
LR (6 W)	0.828 (0.759–0.897)	1.000	0.57	0.602 (0.543–0.660)	0.157	1.000	0.271
LR (1 Y)	0.799 (0.738–0.860)	0.623	0.848	0.799 (0.748–0.844)	0.528	0.623	0.589
XGBoost (6 W)	0.985 (0.907–0.731)	1.000	0.907	0.914 (0.866–0.949)	0.485	1.000	0.653
XGBoost (1 Y)	0.931 (0.806–0.935)	0.957	0.814	0.849 (0.791–0.895)	0.616	0.957	0.750

LR (6 W): logistic regression model predicts rebleeding within 6 weeks; LR (1 Y): logistic regression model predicts rebleeding within 1 year; XGBoost (6 W): XGBoost model predicts rebleeding within 6 weeks; XGBoost (1 Y): XGBoost model predicts rebleeding within 1 year.