Improving feature selection performance using pairwise pre-evaluation

BMC Bioinformatics

Table 5 Comparison of the classification accuracy using the original Relief and the proposed method

Relief		GDS1027		GDS2545		GDS2546		GDS2547		GDS3715
Relief		Orig	Modi	Orig	Modi	Orig	Modi	Orig	Modi	Orig	Modi
5	KNN	0.39	0.52	0.59	0.65	0.56	0.56	0.45	0.59	0.71	0.72
5	SVM	0.46	0.71	0.65	0.72	0.62	0.62	0.55	0.67	0.77	0.83
10	KNN	0.45	0.51	0.59	0.69	0.54	0.56	0.43	0.57	0.72	0.80
10	SVM	0.66	0.73	0.72	0.72	0.66	0.71	0.59	0.62	0.81	0.82
15	KNN	0.47	0.53	0.57	0.71	0.59	0.57	0.46	0.58	0.82	0.84
15	SVM	0.76	0.77	0.71	0.74	0.62	0.74	0.60	0.62	0.83	0.84
20	KNN	0.47	0.55	0.61	0.70	0.56	0.65	0.51	0.57	0.76	0.84
20	SVM	0.81	0.80	0.71	0.74	0.66	0.74	0.66	0.68	0.82	0.86
25	KNN	0.44	0.54	0.56	0.70	0.55	0.60	0.57	0.59	0.73	0.88
25	SVM	0.82	0.87	0.71	0.76	0.66	0.72	0.68	0.68	0.83	0.90
30	KNN	0.45	0.58	0.56	0.73	0.60	0.60	0.59	0.57	0.78	0.86
30	SVM	0.84	0.88	0.70	0.76	0.66	0.70	0.70	0.68	0.85	0.88
MAX KNN		0.47	0.58	0.61	0.73	0.60	0.65	0.59	0.59	0.82	0.88
MAX SVM		0.84	0.88	0.72	0.76	0.66	0.74	0.70	0.68	0.85	0.90

Orig Original algorithm, Modi Proposed modified algorithm
Values in the first column are presented as the number of features selected for the classification test and the others are presented as classification accuracy. The bold numbers denote the highest value of KNN and SVM of each column

ISSN: 1471-2105