Skip to main content

Table 4 The CV results on the rat cirrhosis data

From: Recursive SVM feature selection and sample classification for mass-spectrometry and microarray data

Levela

R-SVM

SVM-RFE

 

CV2b

AveSV c

CV2b

AveSV c

93

4.2%

14.75

4.2%

14.75

80

4.2%

11.91

4.2%

14.74

70

4.2%

9.95

4.2%

14.73

60

3.2%

9.22

4.2%

13.91

50

3.2%

9.03

4.2%

13.82

40

3.2%

9.02

4.2%

14.65

30

3.2%

8.95

4.2%

13.65

20

3.2%

8.93

4.2%

9.98

18

4.2%

8.14

4.2%

9.97

16

4.2%

8.08

3.2%

7.26

15

4.2%

7.60

3.2%

7.15

14

4.2%

7.54

3.2%

7.94

13

6.3%

7.58

4.2%

7.98

12

6.3%

7.41

4.2%

8.05

11

6.3%

7.65

4.2%

8.02

10

6.3%

7.64

3.2%

9.83

9

5.3%

6.50

3.2%

8.83

8

4.2%

5.97

4.2%

7.01

7

4.2%

6.73

4.2%

6.05

6

4.2%

5.98

3.2%

5.97

5

5.3%

5.94

4.2%

5.05

  1. a Level: The number of features selected in each recursive step.
  2. b CV2: Total cross-validation error rate (CV2 error rate).
  3. c AveSV: Average number of support vectors used in the cross-validations at each level.