Table 3.

The number of selected attributes and the classification accuracy under the SVM and KNN classifiers on the raw data and the reduced data with Algorithm 1.

Data Sets	Raw Data			Reduced Data using Algorithm 1			δ
	Attributes	SVM	KNN	Attributes	SVM	KNN
Ionosphere	33	0.874	0.857	11	0.909	0.893	0.3
Wdbc	31	0.538	0.896	10	0.959	0.959	0.15
Wine	13	0.401	0.69	7	0.959	0.96	0.15
Wpbc	33	0.667	0.752	6	0.772	0.753	0.2
Brain_Tumor1	5920	0.86	0.783	13	0.83	0.897	0.15
DLBCL	2000	0.965	0.896	10	0.993	0.998	0.05
Colon	5469	0.811	0.776	5	0.808	0.818	0.15
Lung	7129	0.979	0.975	6	0.99	0.99	0.1
Leukemia	12,533	0.973	0.842	6	0.967	0.981	0.3
Prostate	12,600	0.916	0.796	3	0.829	0.858	0.5
SRBCT	2308	0.984	0.808	6	1	1	0.25
Average	4369.9	0.815	0.825	7.5	0.911	0.919

SVM—support vector machine and KNN—k-nearest neighbors.