Skip to main content

Table 4 The Performance of SVM models using amino acid and PSSM profiles on a realistic dataset (DNArset).

From: Identification of DNA-binding proteins using support vector machines and evolutionary profiles

Threshold

Amino acid composition based model

PSSM based model

 

Sn*

Sp

Acc

MCC

Sn

Sp

Acc

MCC

-1.00

80.21

63.67

65.13

0.26

91.75

77.47

78.73

0.44

-0.90

77.47

67.00

67.92

0.26

89.70

78.60

79.58

0.44

-0.80

76.78

70.67

71.20

0.29

87.66

80.20

80.86

0.45

-0.70

74.74

74.40

74.43

0.31

85.61

81.60

81.95

0.45

-0.60

74.05

77.07

76.79

0.33

84.23

83.20

83.29

0.46

-0.50

71.98

79.27

78.62

0.34

78.73

84.20

83.72

0.44

-0.40

69.91

81.13

80.13

0.34

78.05

85.67

84.99

0.46

-0.30

69.22

83.60

82.32

0.37

74.60

87.27

86.15

0.46

-0.20

65.79

85.73

83.96

0.38

73.22

89.07

87.66

0.48

-0.10

60.99

87.40

85.06

0.37

71.15

91.00

89.24

0.51

0.00

58.23

89.13

86.39

0.38

70.46

92.27

90.34

0.53

0.10

53.40

90.40

87.12

0.37

68.41

93.60

91.37

0.55

0.20

50.02

92.20

88.46

0.38

65.68

94.60

92.04

0.56

0.30

47.95

93.33

89.31

0.40

63.61

95.40

92.59

0.57

0.40

44.53

93.93

89.55

0.38

60.16

95.87

92.71

0.56

0.50

41.79

94.60

89.91

0.38

53.31

96.60

92.77

0.53

0.60

37.70

95.13

90.04

0.35

51.29

97.07

93.01

0.54

0.70

32.90

95.73

90.16

0.33

47.86

97.53

93.13

0.52

0.80

30.83

96.40

90.58

0.33

42.39

97.87

92.95

0.49

0.90

30.16

97.00

91.07

0.35

38.30

98.40

93.07

0.48

1.00

28.78

97.33

91.25

0.35

34.87

98.67

93.01

0.47

  1. * Sn: Sensitivity; Sp: Specificity; Acc: Accuracy