Table 2.

Prediction results of the CSP based feature extraction method and traditional feature extraction methods from evolutionary information.

Method	Sensitivity	Specificity	Accuracy	MCC	AUC	Feature Number
PSSM-DC	0.843	0.774	0.809	0.619	0.873	400
CSP-PSSM-DC	0.705	0.899	0.802	0.615	0.855	20
Bi-gram PSSM	0.710	0.922	0.816	0.646	0.909	400
CSP-Bi-gram PSSM	0.843	0.770	0.806	0.615	0.881	20
ED-PSSM	0.876	0.820	0.848	0.697	0.903	400
CSP-ED-PSSM	0.848	0.829	0.839	0.678	0.908	20