Effective hybrid feature selection using different bootstrap enhances cancers classification performance

Table 18 The proposed methods compared with the Chi-square method

Datasets	Train Data %	Test Data %	Over-fitting Diff. %	Pre	Rec	F1-score	NO.F	F-Time (sec)	C-Time (sec)	AUC	Var.	ACC %
SVM classifier
RNA gene	100.000	99.625	0.375	0.997	0.995	0.996	7555	0.0801	2.379	1.000	0.000036	99.625
DNA CNV	79.862	70.130	9.732	0.592	0.586	0.584	5555	0.528	3.050	0.901	0.000369	70.130
Parkinson’s disease	75.661	72.228	3.433	0.471	0.497	0.448	398	0.016	0.210	0.628	0.000814	72.228
Dermatology diseases	71.220	70.488	0.732	0.556	0.653	0.565	24	0.094	0.093	0.653	0.001305	70.488
BreastEW	91.994	91.563	0.431	0.929	0.895	0.906	21	0.016	0.016	0.976	0.001014	91.563
RF classifier
RNA gene	100.000	99.502	0.498	0.997	0.995	0.996	7555	0.0801	1.009	1.000	0.000041	99.502
DNA CNV	86.934	68.552	18.382	0.585	0.572	0.570	5555	0.528	2.817	0.891	0.000240	68.552
Parkinson’s disease	100.000	81.087	18.913	0.755	0.701	0.704	398	0.016	0.471	0.836	0.008783	81.087
Dermatology diseases	100.000	98.355	1.645	0.984	0.981	0.982	24	0.094	0.229	0.998	0.000363	98.355
BreastEW	100.000	96.832	3.168	0.973	0.962	0.965	21	0.016	0.104	0.990	0.001265	96.832

ISSN: 1756-0381