Skip to main content

Table 3 Overview of precision and recall of all analyzed feature selection methods for different simulation settings

From: Cost-Constrained feature selection in binary classification: adaptations for greedy forward selection and genetic algorithms

 

Recall

Precision

 

A

B

C

D

E

F

G

H

K

A

B

C

D

E

F

G

H

K

Methods

FS

33.9

58.3

19.2

53.7

79.4

34.4

9.7

24.0

18.9

98.7

98.3

98.8

100.0

69.9

93.9

59.1

91.2

98.1

cFS

41.6

66.7

29.1

62.7

69.3

40.2

20.5

27.9

28.2

95.5

100.0

92.5

93.1

37.9

68.2

58.7

85.7

92.3

cFS.mean

38.8

65.7

25.2

59.3

78.3

40.3

18.6

27.0

24.7

97.8

99.0

97.4

97.8

59.8

89.9

65.8

92.0

97.5

cFS.max

37.6

64.0

23.6

57.0

79.1

38.7

16.3

26.3

22.9

98.7

99.0

98.6

98.8

63.5

91.1

65.6

93.0

97.9

fGA

40.6

66.3

27.4

55.0

81.7

38.0

19.0

27.8

23.8

97.9

99.5

98.1

87.3

59.3

83.7

68.1

93.8

98.3

cGA

40.3

66.3

27.3

61.7

79.9

38.6

19.0

27.8

23.8

97.8

99.5

97.0

96.9

61.2

90.0

69.7

94.2

97.4

Filter.tTest

34.3

58.7

18.4

53.0

98.4

34.7

8.8

23.9

18.9

99.5

99.4

99.8

100.0

88.2

95.6

61.5

93.6

99.8

Filter.Symuncert

24.1

56.0

18.4

47.0

84.2

33.0

8.7

24.1

18.6

99.3

100.0

99.6

97.2

84.0

95.0

57.9

99.3

99.8

Filter.PraznikJMIM

30.3

57.3

18.6

44.0

60.8

28.1

7.0

23.8

18.4

83.1

86.0

97.6

76.3

54.4

77.0

43.3

89.3

96.0

Filter.RangerImpurity

34.3

57.0

19.1

53.7

96.0

33.4

8.9

24.1

18.8

93.5

85.5

99.0

98.8

85.0

93.3

60.5

89.9

98.3

Reference

Budget constraint

50.0

66.7

33.3

66.7

100.0

50.0

33.3

33.3

33.3

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

Random selection

20.5

4.7

1.5

0.3

1.1

0.5

1.6

4.0

1.6

60.0

10.0

10.0

1.0

1.0

1.3

10.0

10.0

10.0

  1. Values are given in percent. The results are an extended version of the data shown in Fig. 4. Please refer to the description of this figure for further details