Table 4.

Predictive metrics of different feature selection methods.

Feature selection methods	Optimal subsets	ACC	Recall	MCC	Precision	F-measure
LightGBM	351	0.7360	0.7223	0.4708	0.7240	0.7232
ReliefF	2314	0.7319	0.7150	0.4626	0.7211	0.7181
LinearSVR	2107	0.7304	0.7114	0.4594	0.7204	0.7159
XGBoost	563	0.7351	0.7231	0.4692	0.7224	0.7228
ANOVA	1203	0.7316	0.7142	0.4620	0.7211	0.7176
t-SNE	3	0.5526	0.5112	0.1019	0.5328	0.5218
PCA	1517	0.6952	0.6617	0.3885	0.6880	0.6746