Table 2 Five-fold cross-validation and Independent evaluation (IE) test results of the SVM method for spermatogenesis 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

0.03

82.15

82.12

82.59

64.28

85.99

85.12

85.12

71.88

2

0.03

85.17

81.92

88.02

70.53

84.05

84.3

82.26

68.04

3

0.04

82.73

81.54

83.94

65.47

84.44

84.3

82.93

68.8

4

0.03

83.61

82.88

84.51

67.23

88.33

86.78

88.24

76.56

5

0.05

84.1

81.92

86.06

68.29

81.71

80.99

80.33

63.31

Average

0.04

83.55

82.07

85.02

67.16

84.9

84.29

83.77

69.71

  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.