A New Ensemble Method with Feature Space Partitioning for High-Dimensional Data Classification

Table 8

Paired -test between proposed method and bagging.

	Mean	Standard deviation	Standard error mean	95% confidence interval of the difference			df	Sig.
				Lower	Upper
Proposed method – bagging on AR10P dataset	1.536	1.495	0.451	0.531	2.541	3.407	10	0.007
Proposed method – forest on ORL10P dataset	3.500	1.179	0.373	2.657	4.343	9.391	9	0.000
Proposed method – bagging on PIE10P dataset	1.048	0.770	0.244	0.497	1.599	4.303	9	0.002
Proposed method – bagging on PIX10P dataset	2.100	0.568	0.180	1.694	2.506	11.699	9	0.000
Proposed method – bagging on Leukemia dataset	4.445	1.432	0.453	3.421	5.469	9.816	9	0.000
Proposed method – bagging on Prostate dataset	2.940	1.033	0.327	2.201	3.679	9.000	9	0.000