Table 4.

The classification performance of different modelling methods

Method	Training set (mean ± std)						Test set
	bACC	AUC	SENS	SPEC	PPV	NPV	bACC	AUC	SENS	SPEC	PPV	NPV
Original imbalanced training set
SVM	0.66 ± 0.05	0.76 ± 0.07	0.32 ± 0.11	0.99 ± 0.02	0.93 ± 0.07	0.82 ± . 0.02	0.68	0.78	0.36	1.00	1.00	0.82
AdaBoost	0.76 ± 0.14	0.72 ± 0.16	0.68 ± 0.18	0.84 ± 0.09	09 ± . 0.09	0.89 ± 0.05	0.73	0.79	0.55	0.91	0.68	0.85
Cost-sensitive SVM	0.66 ± 0.15	0.73 ± 0.10	0.43 ± 0.21	0.89 ± 0.09	0.61 ± 0.17	0.84 ± 0.04	0.65	0.75	0.45	0.84	0.49	0.82
Balanced training set augmented with ADASYN
SVM	0.71 ± 0.17	0.79 ± 0.10	0.67 ± 0.17	0.74 ± 0.11	0.45 ± 0.05	0.88 ± 0.04	0.76	0.85	0.73	0.78	0.53	0.89
AdaBoost	0.66 ± 0.14	0.71 ± 0.08	0.55 ± 0.15	0.76 ± 0.07	0.42 ± 0.11	0.85 ± 0.03	0.74	0.82	0.73	0.75	0.5	0.89
Original imbalanced training set (combining data balance and ensemble learning)
SMOTEBoost	0.71 ± 0.11	0.70 ± 0.09	0.52 ± 0.14	0.89 ± 0.10	0.72 ± 0.18	0.85 ± 0.02	0.72	0.80	0.55	0.88	0.61	0.85
RUSBoost	0.77 ± 0.10	0.83 ± 0.13	0.72 ± 0.15	0.82 ± . 0.12	0.74 + 0.02	0.82 ± 0.12	0.83	0.87	0.82	0.84	0.64	0.93

The best results of each metric are shown in bold, and the worst results are shown in italics. Performance evaluation results obtained by bootstrap K-fold cross-validation in the training set