Skip to main content

Table 8 Performance measures of data mining algorithm at different levels of significance on B & D conditions

From: Comparative study of classification algorithms for immunosignaturing data

SIGNIFICANCE

p < 5 x 10-4

p < 5 x 10-3

p < 5 x 10-2

 

Algorithm

Acc.

Sp

Sn

AUC

Acc.

Sp

Sn

AUC

Acc.

Sp

Sn

AUC

Avg.

Naïve Bayes

91.7

100

83.3

0.95

91.7

91.7

91.7

0.92

95.8

91.7

100

0.98

93.6

SVM

91.7

100

83.3

0.92

91.7

91.7

91.7

0.92

95.8

100

91.7

0.96

93.1

VFI

87.5

100

75.0

0.93

91.7

100

83.3

0.94

95.8

100

91.7

1.00

92.7

Logistic R.

79.1

83.3

75.0

0.92

100

100

100

1.00

87.5

91.7

83.3

0.97

90.7

MLP

87.5

91.7

83.3

0.94

87.5

83.3

91.7

0.96

dnf

dnf

dnf

dnf

89.3*

K means

87.5

91.7

83.3

0.88

91.4

91.7

91.7

0.92

87.5

83.3

91.7

0.88

89.0

Hyper Pipes

87.5

83.3

91.7

0.89

91.7

91.7

91.7

0.87

83.3

75.0

91.7

0.90

87.8

Bayes Net

83.3

83.3

83.3

0.89

87.5

91.7

83.3

0.86

83.3

83.3

83.3

0.84

85.1

SLR

83.3

83.3

83.3

0.88

79.2

66.7

91.7

0.90

87.5

100

75.0

0.89

84.7

KNN

79.2

75.0

83.3

0.80

83.3

83.3

83.3

0.83

87.5

91.7

83.3

0.90

83.6

Random Forest

83.3

83.3

83.3

0.83

79.2

83.3

75.0

0.84

79.2

83.3

75.0

0.81

81.1

M5P

87.5

91.7

83.3

0.88

79.2

83.3

75.0

0.73

75.0

83.3

66.7

0.69

79.6

ASC

91.7

100

83.3

0.83

75.0

83.3

66.7

0.61

70.8

75.0

66.7

0.64

76.7

J48

91.7

100

83.3

0.83

75.0

83.3

66.7

0.61

70.8

75.0

66.7

0.64

76.7

Random Tree

83.3

91.7

75.0

0.83

70.8

66.7

75.0

0.71

70.8

66.7

75.0

0.71

75.0

K star

70.8

66.7

75.0

0.83

79.2

75.0

83.3

0.82

58.3

100

16.7

0.58

70.7

LDA

62.5

72.3

60.9

0.75

50.0

65.0

48.0

0.71

20.8

42.6

18.6

0.45

52.6

  1. Acc: Accuracy, Sp: Specificity, Sn: Sensitivity, AUC: Area under ROC curve, Avg: Average score in % for each algorithms, dnf: Did not Finish”, * denotes Avg. from 3 significance levels. Measures >90% are marked in bold.