Effective hybrid feature selection using different bootstrap enhances cancers classification performance

Table 12 The proposed methods compared with the CfsSubsetEval 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	97.125	2.029	0.973	0.974	0.972	4083	3.860	1.030	0.987	0.000627	97.125
DNA CNV	64.034	63.682	0.352	0.539	0.537	0.533	41	2.950	0.009	0.815	0.000997	63.682
Parkinson’s disease	87.683	78.421	9.262	0.730	0.673	0.691	119	0.180	0.016	0.713	0.005994	78.421
Dermatology diseases	74.438	73.776	0.662	0.533	0.598	0.547	9	0.150	0.0002	0.884	0.002967	73.776
BreastEW	94.083	90.865	3.218	0.905	0.902	0.902	3	0.050	0.0002	0.950	0.000596	90.865
Naïve base classifier
RNA gene	99.861	98.503	1.358	0.986	0.979	0.981	4083	3.860	0.048	0.987	0.000131	98.503
DNA CNV	64.864	64.196	0.668	0.601	0.620	0.598	41	2.950	0.001	0.884	0.001219	64.196
Parkinson’s disease	80.904	79.915	0.989	0.753	0.681	0.695	119	0.180	0.001	0.762	0.005410	79.915
Dermatology diseases	62.360	61.441	0.919	0.573	0.628	0.561	9	0.150	0.0007	0.935	0.005554	61.441
BreastEW	93.556	92.973	0.583	0.932	0.921	0.924	3	0.050	0.0008	0.979	0.000886	92.973
K-nearest neighbors(KNN) classifier
RNA gene	99.792	99.627	0.165	0.998	0.996	0.997	4083	3.860	0.007	1.000	0.000036	99.627
DNA CNV	78.986	72.599	6.387	0.691	0.626	0.630	41	2.950	0.00009	0.862	0.000289	72.599
Parkinson’s disease	89.021	80.282	8.739	0.751	0.694	0.708	119	0.180	0.0008	0.743	0.002007	80.282
Dermatology diseases	85.247	79.767	5.48	0.823	0.783	0.780	9	0.150	0.0004	0.943	0.006681	79.767
BreastEW	87.503	81.172	6.331	0.815	0.793	0.797	3	0.050	0.0008	0.857	0.005639	81.172

ISSN: 1756-0381