Skip to main content

Table 3 The results of different gene selection methods over different datasets based on SVM classifier

From: A graph-based gene selection method for medical diagnosis problems using a many-objective PSO algorithm

Dataset

 

Gene selection method

MaPSOGS

Geometric PSO

PSO

RMA

RPSW

EPSO

Hybrid BPSO-BBHA

PSOC4.5

AMLGSE2191

Acc (%)

100 (1)

89.01 (7)

82.29 (8)

96.29 (3)

90.48 (6)

91.63 (5)

94.26 (4)

96.88 (2)

 

Std

1.58

2.68

3.54

1.80

2.11

2.37

3.01

2.93

Colon

Acc (%)

99.52 (2)

89.93 (7)

86.63 (8)

100 (1)

91.38 (5)

91.01 (6)

92.25 (4)

95.27 (3)

 

Std

1.90

2.25

3.18

2.81

3.18

3.20

1.31

6.38

DLBCL

Acc (%)

98.84 (1)

86.08 (7)

83.54 (8)

95.15 (2)

94.24 (3)

91.75 (6)

92.06 (5)

92.65 (4)

 

Std

1.36

0.63

0.81

0.88

2.32

1.20

0.77

1.65

Leukaemia

Acc (%)

98.71 (1)

86.13 (7)

82.39 (8)

96.81 (2)

91.90 (4)

90.96 (5)

89.92 (6)

93.46 (3)

 

Std

1.90

1.44

1.85

1.44

1.71

1.58

0.76

1.39

Prostate

Acc (%)

98.38 (1)

85.96 (7)

84.55 (8)

97.41 (2)

90.51 (4)

89.14 (5)

87.92 (6)

91.81 (3)

 

Std

0.33

2.10

1.60

1.29

1.21

1.60

2.76

4.51

MLL

Acc (%)

96.38 (1)

81.26 (8)

81.67 (7)

91.23 (3)

86.53 (5)

85.98 (6)

86.75 (4)

91.63 (2)

 

Std

1.51

3.46

4.26

3.63

3.40

3.64

4.65

3.32

SRBCT

Acc (%)

98.91 (1)

84.56 (7)

83.28 (8)

95.37 (2)

91.50 (5)

90.11 (6)

91.70 (4)

93.03 (3)

 

Std

2.42

4.44

2.50

4.87

4.80

1.49

3.63

2.98

Average

Acc (%)

98.68 (1.14)

86.13 (7.14)

83.48 (7.85)

96.04 (2.14)

90.93 (4.57)

90.08 (5.57)

90.69 (4.71)

93.53 (2.85)

 

Std

1.57

2.43

2.53

2.39

2.67

2.15

2.41

3.31

  1. The best results are shown in bold