Effective hybrid feature selection using different bootstrap enhances cancers classification performance

Table 13 The proposed methods compared with the ReliefAttributeEval method

Datasets	Train Data %	Test Data %	Over-fitting Diff. %	Pre	Rec	F1-score	NO.F	F-Time (sec)	C-Time (sec)	AUC	Var.	ACC %
J48 classifier
RNA gene	99.154	97.625	1.529	0.980	0.979	0.979	10,000	1.950	2.887	0.992	0.000432	97.625
DNA CNV	65.322	63.922	1.4	0.546	0.548	0.540	8000	1.550	0.994	0.816	0.001225	63.922
Parkinson’s disease	83.858	74.759	9.099	0.635	0.599	0.593	300	0.950	0.060	0.710	0.013702	74.759
Dermatology diseases	79.356	78.701	0.655	0.570	0.647	0.591	20	0.500	0.0007	0.924	0.000719	78.701
BreastEW	96.485	93.499	2.986	0.467	0.448	0.455	16	0.350	0.003	0.959	0.002994	93.499
Naïve base classifier
RNA gene	99.855	96.881	2.974	0.962	0.955	0.955	10,000	1.950	0.115	0.973	0.000976	96.881
DNA CNV	65.678	64.679	0.999	0.629	0.668	0.620	8000	1.550	0.297	0.849	0.001359	64.679
Parkinson’s disease	81.550	79.338	2.212	0.742	0.729	0.722	300	0.950	0.003	0.775	0.014727	79.338
Dermatology diseases	87.403	85.570	1.833	0.808	0.852	0.806	20	0.500	0.0007	0.979	0.002454	85.570
BreastEW	94.681	94.415	0.266	0.481	0.444	0.460	16	0.350	0.0005	0.989	0.002167	94.415
K-nearest neighbors(KNN) classifier
RNA gene	99.875	99.873	0.002	0.999	0.999	0.999	10,000	1.950	0.016	1.000	0.000031	99.873
DNA CNV	80.735	74.246	6.489	0.708	0.655	0.654	8000	1.550	0.013	0.874	0.000378	74.246
Parkinson’s disease	85.801	71.708	14.093	0.595	0.586	0.579	300	0.950	0.0007	0.608	0.007203	71.708
Dermatology diseases	92.289	86.059	6.23	0.872	0.848	0.840	20	0.500	0.001	0.961	0.004704	86.059
BreastEW	94.005	91.739	2.266	0.462	0.427	0.442	16	0.350	0.0002	0.964	0.000553	91.739

ISSN: 1756-0381