A Comparison among Different Machine Learning Pretest Approaches to Predict Stress-Induced Ischemia at PET/CT Myocardial Perfusion Imaging

Table 3

Metrics obtained from the ML techniques, evaluated on training/test and validation approaches.

	Training/test ()				Validation ()
	Accuracy (%)	Sensitivity (%)	Specificity (%)	AUROC (%)	Accuracy (%)	Sensitivity (%)	Specificity (%)	AUROC (%)
ADA	88	48	97	90	76	26	89	68
AdaBoost	89	67	95	95	71	23	87	66
Logistic	80	5	98	72	80	7	98	75
Naïve Bayes	77	23	91	70	80	27	92	73
Random Forest	89	51	98	93	75	21	89	65
Rpart	82	27	96	75	76	17	91	70
SVM	72	13	87	61	77	21	91	65
XGBoost	83	27	97	83	77	18	92	69