Skip to main content

Table 4 The results of different gene selection methods over different datasets based on DT 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 (%)

97.08 (1)

86.65 (7)

80.90 (8)

95.11 (2)

91.18 (5)

90.01 (6)

91.63 (4)

94.05 (3)

 

Std

5.08

4.45

6.16

3.96

4.39

3.88

4.24

2.77

Colon

Acc (%)

96.52 (1)

87.05 (7)

84.09 (8)

94.89 (2)

91.55 (4)

89.14 (6)

90.95 (5)

92.64 (3)

 

Std

2.29

3.65

4.02

2.07

4.78

3.95

2.47

3.95

DLBCL

Acc (%)

94.57 (1)

85.62 (7)

81.33 (8)

92.63 (2)

91.07 (3)

90.22 (5)

89.96 (6)

90.91 (4)

 

Std

1.54

5.72

6.16

2.85

4.66

3.39

7.05

4.48

Leukaemia

Acc (%)

95.42 (1)

83.39 (7)

82.12 (8)

92.89 (2)

89.69 (4)

88.43 (6)

88.53 (5)

91.57 (3)

 

Std

0.86

4.17

4.55

1.01

2.33

3.97

3.91

1.93

Prostate

Acc (%)

96.71 (1)

81.93 (8)

83.05 (7)

95.09 (2)

90.28 (4)

86.14 (6)

87.19 (5)

91.18 (3)

 

Std

3.75

4.99

7.87

2.05

5.66

2.84

5.41

2.97

MLL

Acc (%)

93.44 (1)

80.98 (7)

79.24 (8)

90.93 (2)

85.87 (5)

89.88 (4)

85.28 (6)

90.22 (3)

 

Std

2.12

5.83

5.75

4.16

2.49

4.18

6.17

3.63

SRBCT

Acc (%)

95.85 (1)

81.66 (8)

80.88 (7)

92.16 (2)

90.02 (5)

90.75 (4)

89.97 (6)

91.76 (3)

 

Std

4.09

7.02

6.58

3.90

5.54

5.40

4.11

3.75

Average

Acc (%)

95.65 (1)

83.89 (7.28)

81.66 (7.71)

93.38 (2)

89.95 (4.28)

89.22 (5.28)

89.07 (5.28)

91.76 (3.14)

 

Std

2.82

5.12

5.87

2.86

4.26

3.94

4.76

3.35

  1. The best results are shown in bold