Skip to main content

Table 3 The experimental results of the DAEGCNDF model based on 10-fold cross-validation

From: Predicting potential microbe-disease associations based on auto-encoder and graph convolution network

Dataset:

MDAID

Testing set

Acc(%)

Pre (%)

Recall (%)

F1-score (%)

AUC (%)

AUPR (%)

1

92.34

92.35

92.33

92.34

97.71

98.01

2

90.68

90.62

90.68

90.65

97.42

97.12

3

90.01

90.03

90.01

90.01

96.65

96.37

4

89.90

89.89

89.86

89.87

96.04

95.65

5

91.01

90.97

91.05

91.00

97.42

97.61

6

91.79

91.79

91.79

91.79

97.55

97.71

7

90.12

90.11

90.19

90.12

95.87

95.36

8

89.67

89.65

89.69

89.66

96.36

96.40

9

90.22

90.34

90.13

90.19

97.03

96.87

10

93.11

93.14

93.10

93.11

97.90

97.90

Average

90.89 ± 1.16

90.89 ± 1.17

90.88 ± 1.16

90.87 ± 1.17

97.00 ± 0.72

96.90 ± 0.94

Dataset:

HMDAD

Testing set

Acc(%)

Pre (%)

Recall (%)

F1-score (%)

AUC (%)

AUPR (%)

1

91.11

91.25

90.84

91.00

97.66

97.65

2

90.00

90.00

90.18

89.89

97.07

9755

3

85.56

86.25

85.87

85.54

94.85

94.86

4

86.67

86.84

86.84

86.67

96.14

96.28

5

86.67

87.16

86.94

86.67

95.45

96.30

6

83.33

83.90

83.78

83.33

93.25

94.78

7

92.22

92.16

92.26

92.20

97.17

97.49

8

90.00

90.18

90.00

90.00

97.93

98.11

9

88.89

89.00

88.69

88.80

96.78

96.82

10

86.67

86.74

86.67

86.66

96.94

97.37

Average

88.11 ± 2.77

88.35 ± 2.58

88.21 ± 2.62

88.08 ± 2.75

96.32 ± 1.44

96.72 ± 1.16