Table 3.

Model parameters in validation set.

Model	AUC	Cutoff	Accuracy	Sensitivity	Specificity	PPV	NPV (SD)	F1-Score
XGBoost	0.800 (0.019)	0.463 (0.013)	0.749 (0.017)	0.718 (0.060)	0.790 (0.036)	0.643 (0.035)	0.812 (0.026)	0.677 (0.035)
Logistic	0.773 (0.025)	0.468 (0.033)	0.738 (0.028)	0.650 (0.065)	0.821 (0.064)	0.649 (0.051)	0.788 (0.045)	0.648 (0.050)
Light GBM	0.531 (0.081)	0.600 (0.800)	0.626 (0.042)	0.350 (0.308)	0.751 (0.292)	NA	0.651 (0.038)	NA
AdaBoost	0.791 (0.028)	0.467 (0.002)	0.714 (0.031)	0.695 (0.065)	0.782 (0.053)	0.584 (0.056)	0.820 (0.036)	0.633 (0.052)
GNB	0.790 (0.031)	0.351 (0.032)	0.738 (0.037)	0.694 (0.065)	0.807 (0.073)	0.639 (0.066)	0.809 (0.045)	0.664 (0.057)
CNB	0.671 (0.029)	0.993 (0.022)	0.653 (0.031)	0.677 (0.042)	0.661 (0.028)	NA	0.739 (0.066)	NA
MLP	0.606 (0.077)	0.397 (0.029)	0.617 (0.107)	0.548 (0.305)	0.664 (0.289)	NA	NA	NA
SVM	0.524 (0.093)	0.449 (0.057)	0.569 (0.100)	0.554 (0.292)	0.576 (0.316)	0.467 (0.126)	0.717 (0.146)	0.456 (0.155)

Data are shown as means ± standard deviations (SD).

AUC, area under the receiver operating characteristic curve; PPV, positive prediction value; NPV, negative prediction value; GNB, Gaussian Naïve Bayes; CNB, Complement Naive Bayes; GNB, Gaussian Naïve Bayes; MLP, multi-layer perceptron neural network; SVM, support vector machine; NA, not applicable.