Table 1. The performances of SVM model and RF model in classification of MFEs.

	Positives	Negatives	TP	FP	TN	FN	SP (%)	SE (%)	PPV (%)	Q (%)
SVM	6,782	10,714	5,642	1,435	9,279	1,140	86.6	83.2	79.7	85.3
RF	6,782	10,714	6,368	632	10,082	414	94.1	93.9	91.0	94.0

The prediction were evaluated by parameters of TP (true positive), FN (false negative), TN (true negative), FP (false positive), specificity SP =  TN/(TN+FP), sensitivity SE =  TP/(TP+FN), positive prediction value PPV = TP/ (TP+FP) and overall accuracy Q =  (TP+TN)/ (TP+FN+TN+FP).