Feature Selection of the Rich Model Based on the Correlation of Feature Components

Table 2

The performances of the original s1-minmax22h submodel of SRM and the subfeatures selected by different methods.

Payload	(325-D)	Random		Proposed
		Number of feature components		Number of feature components
0.1	0.5397	82	0.5162	82	0.5391
0.2	0.5873	81	0.5567	81	0.5828
0.3	0.6278	80	0.5432	80	0.6266
0.4	0.6709	84	0.5994	84	0.6686
0.5	0.7053	81	0.6274	81	0.7046