A genetic algorithm-based weighted ensemble method for predicting transposon-derived piRNAs

BMC Bioinformatics

Table 8 Performances of GA-WE and the state-of-the-art methods on three species

Dataset	Species	Method	AUC	ACC	SN	SP
Balanced	Human	Piano	0.592	0.560	0.855	0.265
		piRNApredictor	0.894	0.812	0.859	0.764
		Ensemble Learning	0.920	0.807	0.815	0.800
		GA-WE	0.932	0.839	0.858	0.820
	Mouse	Piano	0.445	0.5365	0.837	0.236
		piRNApredictor	0.892	0.819	0.862	0.776
		Ensemble Learning	0.924	0.810	0.863	0.756
		GA-WE	0.937	0.838	0.826	0.850
	Drosophila	Piano	0.741	0.692	0.836	0.547
		piRNApredictor	0.983	0.952	0.927	0.977
		Ensemble Learning	0.994	0.958	0.952	0.965
		GA-WE	0.995	0.959	0.949	0.966
Imbalanced	Human	Piano	0.449	0.747	0.000	1.000
		piRNApredictor	0.905	0.847	0.548	0.949
		Ensemble Learning	0.922	0.836	0.589	0.919
		GA-WE	0.935	0.869	0.687	0.931
	Mouse	Piano	0.441	0.744	0.000	1.000
		piRNApredictor	0.892	0.848	0.568	0.944
		Ensemble Learning	0.928	0.849	0.586	0.940
		GA-WE	0.939	0.889	0.745	0.939
	Drosophila	Piano	0.804	0.712	0.000	1.000
		piRNApredictor	0.982	0.961	0.902	0.985
		Ensemble Learning	0.995	0.965	0.920	0.984
		GA-WE	0.996	0.964	0.940	0.973

ISSN: 1471-2105