Table 4.

The results of 10-fold cross-validation by the five methods averaged over 13 classes

Methods	precision(%)	recall(%)	F1(%)
AGPS	68	61	61
PSoL	68	37	47
two-class SVMs	45	24	33
two-class SVMs_balanced	61	70	69
one-class SVMs	50	21	31
kernel integration	58	28	37
kernel integration_balanced	64	47	52