Table 3.

The performance of models trained with combined features.

Training Feature	Sensitivity	Specificity	Accuracy	MCC	AUC
3-gap DC+CSP-PSSM-DC	0.853	0.816	0.834	0.669	0.887
3-gap DC+CSP-Bi-gram PSSM	0.853	0.839	0.846	0.691	0.887
3-gap DC+CSP-ED-PSSM	0.876	0.843	0.859	0.719	0.905
3-gap DC+CSP-PSSM-DC+CSP-Bi-gram PSSM	0.857	0.793	0.825	0.651	0.882
3-gap DC+CSP-PSSM-DC+CSP-ED-PSSM	0.862	0.843	0.853	0.705	0.899
3-gap DC+CSP-Bi-gram PSSM+CSP-ED-PSSM	0.843	0.839	0.841	0.682	0.894
3-gap DC+CSP-PSSM-DC+CSP-Bi-gram PSSM+CSP-ED-PSSM	0.876	0.853	0.864	0.728	0.912