Skip to main content

Table 5 The results of different gene selection methods over different datasets based on KNN 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 (%)

95.36 (1)

87.08 (7)

81.16 (8)

95.11 (2)

90.79 (6)

91.22 (4)

91.05 (5)

93.39 (3)

 

Std

2.33

7.15

5.78

1.23

8.04

3.99

4.57

3.50

Colon

Acc (%)

96.18 (1)

86.76 (7)

84.88 (8)

94.10 (2)

90.55 (4)

88.21 (6)

90.36 (5)

91.45 (3)

 

Std

0.24

4.77

6.42

1.70

5.49

4.87

5.12

3.10

DLBCL

Acc (%)

93.89 (1)

86.96 (7)

82.83 (8)

91.49 (2)

91.18 (3)

89.64 (5)

88.07 (6)

90.39 (4)

 

Std

0.59

5.11

4.96

1.94

3.07

2.79

2.94

2.28

Leukaemia

Acc (%)

93.81 (1)

84.14 (7)

82.66 (8)

91.95 (3)

88.71 (5)

87.19 (6)

89.06 (4)

91.88 (2)

 

Std

4.17

4.17

3.69

2.91

4.17

3.60

4.02

3.62

Prostate

Acc (%)

94.25 (1)

80.83 (8)

82.69 (7)

93.77 (2)

90.26 (3)

85.55 (6)

86.37 (5)

90.04 (4)

 

Std

2.41

5.53

7.01

4.84

4.26

5.03

2.50

3.09

MLL

Acc (%)

93.74 (1)

81.49 (7)

80.90 (8)

91.55 (2)

83.72 (6)

88.78 (4)

86.44 (5)

90.61 (3)

 

Std

1.01

3.13

4.25

3.05

3.83

4.45

3.96

3.77

SRBCT

Acc (%)

94.90 (1)

81.96 (7)

81.13 (8)

92.53 (2)

90.11 (4)

90.01 (5)

88.83 (6)

91.39 (3)

 

Std

2.85

4.18

3.18

3.21

5.73

3.97

2.98

2.19

Average

Acc (%)

94.59 (1)

84.18 (7.14)

82.32 (7.85)

92.93 (2.14)

89.33 (4.42)

88.66 (5.14)

88.60 (5.14)

91.31 (3.14)

 

Std

1.94

4.86

5.04

2.70

4.94

4.1

3.73

3.08

  1. The best results are shown in bold