Sequence-based bacterial small RNAs prediction using ensemble learning strategies

BMC Bioinformatics

Table 7 Performances of different methods on benchmark datasets

Dataset	Ratio	Method	AUC	ACC	SN	SP
Balanced	1:1	Carter’s method	0.566	0.511	0.264	0.758
		Barman’s method	0.938	0.882	0.846	0.918
		WAEM	0.942	0.887	0.888	0.868
		NNEM	0.958	0.901	0.903	0.899
Imbalanced	1:2	Carter’s method	0.602	0.678	0.033	1.000
		Barman’s method	0.937	0.884	0.851	0.916
		WAEM	0.952	0.901	0.853	0.925
		NNEM	0.962	0.909	0.872	0.927
	1:3	Carter’s method	0.619	0.757	0.030	1.000
		Barman’s method	0.944	0873	0.818	0.927
		WAEM	0.951	0.915	0.818	0.948
		NNEM	0.961	0.920	0.819	0.954
	1:4	Carter’s method	0.627	0.805	0.025	1.000
		Barman’s method	0.944	0.874	0.818	0.929
		WAEM	0.957	0.929	0.817	0.956
		NNEM	0.962	0.931	0.810	0.961
	1:5	Carter’s method	0.636	0.835	0.011	1.000
		Barman’s method	0.943	0.875	0.884	0.865
		WAEM	0.957	0.934	0.808	0.959
		NNEM	0.961	0.940	0.782	0.972

ISSN: 1471-2105