Skip to main content

Table 1 Summary results for the SVM-RNE, SVM-RCE and SVM-RFE algorithms.

From: Classification and biomarker identification using gene network modules and support vector machines

 

CTCL(I)

CTCL(II)

Lymphocyte

 

c

g

ACC

c

g

ACC

c

g

ACC

SVM-RNE

2

4

100%

2

5

91%

4

13

80%

 

4

8

97%

4

8

90%

6

18

79%

 

14

31

96%

14

33

90%

11

30

74%

 

24

55

92%

30

69

89%

   

SVM-RCE

2

8

96%

2

8

76%

2

13

96%

 

3

12

96%

9

34

89%

   
 

9

32

97%

19

71

91%

   
 

15

51

97%

28

104

91%

   
 

32

101

96%

      
       

6

39

92%

       

10

64

92%

SVM-RFE

 

9

89%

 

8

84%

 

12

81%

  

32

94%

 

32

85%

 

18

79%

  

102

100%

 

102

87%

 

30

77%

  1. Classification accuracies for the three algorithms on three datasets are presented at representative steps in the course of recursive feature elimination. The number of clusters (c) and the total number of genes in the clusters (g) are shown for the steps which are presented in reverse order, i.e. the last elimination step is shown first in the table. No clusters are shown for SVM-RFE since the genes are eliminated without clustering.
Back to article page

BMC Bioinformatics

ISSN: 1471-2105

Contact us