Table 4.

Independent test results for different feature representations using class and probabilistic information.

Features	Dimension	ACC	Sn	Sp	MCC	AUC
PF	130	0.857	0.937	0.778	0.723	0.924
CF	130	0.817	0.746	0.889	0.642	0.892
PCF	260	0.857	0.778	0.937	0.723	0.925
Optimal PF	50	0.881	0.810	0.952	0.770	0.922
Optimal CF	5	0.802	0.794	0.810	0.603	0.859
Optimal PCF	5	0.873	0.841	0.905	0.748	0.891