Table 10. Comparison of the accuracy of the most accurate models (i.e. model 40 for hard90 and model 50 for hard70) with the most accurate models based various FS techniques, and also model 40 with model 50.

The differences between these comparisons based on the Mann-Whitney tests (n = 100 for each model).

FS method Hard90	Models		p-value for ccr					p-value for kappa
	Developed	Model	24	27	30	45	46	24	27	30	45	46
VI & filter	25	24
AVI & filter	4	27	0.8881					0.0010
KIAVI	2+25+4	30	0.0003	0.0006				0.0000	0.4450
Boruta	2	45	0.0000	0.0000	0.0000			0.0000	0.0000	0.0000
GRF	1	46	0.0000	0.0000	0.0000	0.0000		0.0000	0.0000	0.0000	0.0000
AVI	11	40	0.0000	0.0000	0.0000	0.0000	0.0000	0.0000	0.0000	0.0000	0.0000	0.0000
Hard70		Model	24	33	47	53		24	33	47	53
AVI & filter	25	24
AVI	13	33	0.1079					0.3001
KIAVI	11+25+13	47	0.6253	0.1820				0.9357	0.2328
GRF	1	53	0.0000	0.0000	0.0000			0.0000	0.0000	0.0000
Boruta	3	50	0.0000	0.0000	0.0000	0.0000		0.0000	0.0000	0.0000	0.0000
Hard90/Hard70		Model	40					40
		50	0.0000					0.0000