Table 4.

ROC curve analysis of three models in the testing set.

Parameter	Logistic regression	Support vector machine	Decision Tree
AUC	0.639 (CI: 0.476–0.802)	0.826 (CI: 0.688–0.964)	0.717 (CI: 0.564–0.871)
Accuracy	0.596 (CI: 0.458–0.724)	0.825 (CI: 0.701–0.913)	0.772 (CI: 0.642–0.873)
Sensitivity	0.750 (CI: 0.436–0.938)	0.750 (CI: 0.438–0.938)	0.625 (CI: 0.320–0.812)
Specificity	0.537 (CI: 0.195–0.756)	0.854 (CI: 0.341–0.976)	0.829 (CI: 0.400–0.951)

ROC, Receiver operating characteristic; AUC, area under the operating curve; CI, confidence interval.