Skip to main content

Table 2 Test results for accuracy and time of the entire transformation dataset

From: Determining threshold value on information gain feature selection to increase speed and prediction accuracy of random forest

Dataset

Number of Instance

Number of Feature

CBFS Best First

Threshold 0,05

Treshold based on Standard Deviation

Number of Feature

Accuracy

Time

Number of Feature

Accuracy

Time

Number of Feature

Accuracy

Time

EEG Eye

14.980

14

4

77.17%

7.31

3

72.06%

4.96

10

90.14%

9.41

Cancer

569

31

12

95.68%

0.08

26

96.68%

0.10

15

94.41%

0.08

ContraceptiveMethod

1.473

9

4

51.74%

0.35

3

50.90%

0.21

4

51.74%

0.27

Dermatology

366

33

15

97.70%

0.07

32

97.35%

0.07

26

97.40%

0.07

Divorce

170

54

6

96.65%

0.02

54

97.71%

0.02

53

97.65%

0.03

Electrical Grid

10.000

14

9

85.64%

5.06

5

76.73%

2.57

7

80.85%

3.44

CNAE-9

1.080

857

28

81.16%

0.27

57

90.69%

0.76

65

90.49%

0.91

Urban Land Cover

168

148

28

87.62%

0.04

110

85.89%

0.07

65

84.64%

0.05

Epilepsy

11.500

179

119

69.51%

21.52

178

69.73%

27.93

178

69.73%

27.93

SCADI

70

206

16

85.86%

0.02

58

85.00%

0.02

58

85.00%

0.02