Table 1.

Performance of machine learning models for training datasets and testing datasets.

	AUC	Accuracy	Sensitivity (%)	Specificity (%)	PPV (%)	NPV (%)
The training datasets	SVM	99.8	97.6	100	95.2	95.5	100
	MLR	98.0	92.9	95.2	90.5	90.9	95.0
	ADA	94.0	85.7	90.5	81.0	82.6	89.5
The testing datasets	SVM	85.2	77.8	66.7	88.9	85.7	72.7
	MLR	84.0	72.2	66.7	77.8	75.0	70.0
	ADA	68.5	66.7	66.7	66.7	66.7	66.7

SVM, support vector machine; MLR, multivariate logistic regression; ADA, AdaBoost; PPV, positive predictive value; NPV, negative predictive value.