6mA-StackingCV: an improved stacking ensemble model for predicting DNA N6-methyladenine site

Table 8 The performance of different classifiers with different representations

Second-layer classifier	Feature	SN	SP	ACC	MCC	AUC
SVM	One-hot + EIIP	0.9589	0.9614	0.9601	0.9203	0.9734
	One-hot + NCP	0.9551	0.9611	0.9581	0.9162	0.9707
	One-hot + ENAC	0.9585	0.9597	0.9591	0.9182	0.9696
	One-hot + EIIP + NCP	0.9542	0.9621	0.9581	0.9163	0.9706
LightGBM	One-hot + EIIP	0.9605	0.9592	0.9599	0.9197	0.9912
	One-hot + NCP	0.9564	0.9603	0.9584	0.9167	0.9907
	One-hot + ENAC	0.9593	0.9575	0.9584	0.9168	0.9909
	One-hot + EIIP + NCP	0.9537	0.9622	0.9579	0.9159	0.9911
logistic regression	One-hot + EIIP	0.9589	0.9611	0.9600	0.9200	0.9913
	One-hot + NCP	0.9563	0.9610	0.9586	0.9173	0.9909
	One-hot + ENAC	0.9600	0.9590	0.9595	0.9190	0.9913
	One-hot + EIIP + NCP	0.9553	0.9607	0.9580	0.9160	0.9909

ISSN: 1756-0381