TABLE 4.

The mean values of seven evaluation metrics obtained from four methods on integrative dataset.

Classifier	B_ACC	ACC	SES	SPC	PRC	F1	AUC
SVM	0.9413	0.9865	0.9910	0.9435	0.9941	0.9926	0.9672
RF	0.9208	0.9902	0.9968	0.9261	0.9924	0.9946	0.9615
KNN	0.9480	0.9914	0.9955	0.9522	0.9950	0.9953	0.9738
Proposed	0.9731	0.9951	0.9964	0.9826	0.9982	0.9973	0.9895

In the experiments, we randomly split the dataset into 10 equal-sized datasets. The mean values of the seven metrics are obtained on the 10 test sets. The proposed method outperforms other methods in balanced accuracy, accuracy, specificity, precision, F1 score, and AUC. The bold values are the best results.