Table 3 Five-fold cross-validation and Independent evaluation (IE) test results of the SVM method for embryogenesis datasets.

From: PrESOgenesis: A two-layer multi-label predictor for identifying fertility-related proteins using support vector machine and pseudo amino acid composition approach

Datasets

λ*

Five-fold cross-validation test

Independent evaluation test

Accuracy (%)

Sensitivity (%)

Specificity (%)

MCC (%)

Accuracy (%)

Sensitivity (%)

Specificity (%)

MCC (%)

1

0.02

80.23

80.74

80.38

45.08

80.12

81.41

77.44

47.33

2

0.03

81.05

79.7

82.39

62.14

79.22

77.56

78.06

58.26

3

0.001

80.75

80.59

81.32

65.2

83.43

78.85

84.83

69.42

4

0.03

82.33

81.19

83.54

67.79

78.92

79.49

76.54

62.23

5

0.001

81.43

81.19

82.04

62.85

80.42

82.05

77.58

60.91

Average

0.02

81.15

80.68

81.93

60.61

80.42

79.87

78.89

59.63

  1. *The optimum λ parameter value of kernel function of SVM using a grid-search technique based on five-fold cross-validation. Also, the optimum parameter C value was obtained 100 in all of models.
Back to article page