Skip to main content

Table 4 Performance comparison of five commonly used binary classifiers on proposed features

From: TNFPred: identifying tumor necrosis factors using hybrid features based on word embeddings

Classifier

Cross-validation data

 

Acc (%)

Spec (%)

Sen (%)

MCC

SVM

95.82 ± 1.67

97.59 ± 2.15

83.67 ± 7.45

0.83 ± 0.06

kNN

77.33 ± 3.7

75.41 ± 3.98

100 ± 0

0.47 ± 0.03

RandomForest

94.22 ± 2.3

94.20 ± 2.9

94 ± 8.43

0.75 ± 0.05

Naïve Bayes

21.59 ± 10.62

14.76 ± 11.45

100 ± 0

0.09 ± 0.06

QuickRBF

94.80 ± 1.52

99.81 ± 0.4

57.99 ± 14.25

0.72 ± 0.09

 

Independent data

 

Acc (%)

Spec (%)

Sen (%)

MCC

SVM

96.49 ± 4.34

98 ± 5.27

85 ± 17.48

0.86 ± 0.13

kNN

79.39 ± 8.9

78.01 ± 10.57

93.34 ± 14.04

0.47 ± 0.09

RandomForest

97.28 ± 2.25

99 ± 2.26

80.01 ± 23.31

0.84 ± 0.14

Naïve Bayes

19.09 ± 23.76

10.99 ± 26.15

100 ± 0

0.08 ± 0.17

QuickRBF

94.12 ± 1.97

100 ± 0

50 ± 16.7

0.68 ± 0.13

  1. (Each result is reported in format: m ± d, where m is the mean and d is the standard deviation across the ten runs)