Gene selection via improved nuclear reaction optimization algorithm for cancer classification in high-dimensional data

Journal of Big Data

Table 35 The number of extracted features by the proposed RBNRO-DE with different filter and embedded methods for training the k-NN

Benchmark	RBNRO-DE	Variance Threshold	Correlation	Chi-square	ANOVA	Linear-Regression	Ridge Regularization	Lasso Regularization	Elastic Net
BLCA	147	500	412	300	300	206	200	310	500
CESC	125	500	439	300	300	209	203	220	500
CHOL	125	500	152	290	280	204	158	180	200
COAD	125	500	287	310	190	202	201	270	200
ESCA	125	500	339	300	300	206	234	250	200
GBM	125	500	254	209	190	199	255	190	500
HNSC	139	500	382	310	300	202	180	420	500
KICH	125	500	106	250	160	196	231	190	160
KIRC	153	500	470	300	300	206	199	390	200
KIRP	125	500	339	290	280	202	199	170	200
LIHC	125	500	410	130	200	218	192	420	200
LUAD	130	500	423	240	300	196	231	430	200
LUSC	131	500	293	300	300	218	194	350	300
PAAD	125	500	360	300	300	205	221	250	500
PCPG	125	500	327	500	500	199	199	200	500
READ	125	500	324	500	500	203	235	190	300
SARC	125	500	431	500	500	200	200	200	500
SKCM	125	500	242	500	500	181	192	180	500
STAD	144	500	428	500	500	217	216	340	200
THCA	125	500	311	500	300	199	194	152	300
THYM	125	500	369	300	300	203	205	210	500
UCEC	151	500	422	500	500	219	206	280	300
Ranking (\(\text {W\|T\|L}\))	\(\mathbf {21\|0\|1}\)	0\|0\|22	1\|0\|22	0\|0\|22	0\|0\|22	0\|0\|22	0\|0\|22	0\|0\|22	0\|0\|22