Table 5.

The performance of all feature descriptors with various machine learning algorithms based on independent dataset

Method	Features	ACC	SN	SP	AUC	MCC	F1
Random forest	OAAC	0.697	0.734	0.585	0.660	0.290	0.785
	Dipeptide = 0	0.570	0.557	0.610	0.583	0.144	0.660
	Dipeptide = 1	0.696	0.731	0.590	0.687	0.292	0.784
	Dipeptide = 2	0.546	0.516	0.634	0.575	0.130	0.631
	AAindex	0.703	0.798	0.415	0.607	0.211	0.802
	PSSM	0.703	0.774	0.488	0.631	0.249	0.797
	All features	0.727	0.807	0.488	0.647	0.288	0.816
SVM	All features	0.642	0.613	0.732	0.672	0.298	0.720
Baseline	All features	0.509	0.492	0.558	0.526	0.044	0.591