Skip to main content
. 2021 Jun 2;11:11591. doi: 10.1038/s41598-021-90991-0

Table 3.

Overfitting evaluation of the prediction models.

Models AUC [95%CI] P-value
Training cohort Test cohort
LR 0.886 [0.817–0.936] 0.845 [0.732–0.924] 0.4426
SVM 0.845 [0.769–0.904] 0.829 [0.713–0.912] 0.8556
DT 0.832 [0.754–0.893] 0.827 [0.711–0.911] 0.7253
RF 0.812 [0.769–0.904] 0.843 [0.729–0.922] 0.5767
XGBoost 0.858 [0.784–0.914] 0.836 [0721–0.917] 0.7146
Clinical model 0.730 [0.642–0.807] 0.805 [0.686–0.894] 0.2854
Combined model 0.921 [0.858–0.962] 0.865 [0.757–0.938] 0.3245

P-value reflected the differences between the training and test cohorts, and P < 0.05 (two-sided) were considered statistically significant.

AUC area under the curve; CI confidence interval; LR logistic regression; SVM support vector machine; DT decision tree; RF random forest; XGBoost extreme gradient boosting.