Skip to main content

Table 4 Fivefold cross-validation performance analysis of deep-learning based machine learning methods on unbalanced dataset (1:3)

From: Struct2Graph: a graph attention network for structure based predictions of protein–protein interactions

Method

Performance (%) on unbalanced training set—1:3

Accuracy

Precision

Recall

Specificity

DeepPPI

98.19 ± 0.58

98.73 ± 0.40

93.98 ± 2.43

99.59 ± 0.13

DeepFE-PPI

98.96 ± 0.27

98.30 ± 0.46

97.52 ± 0.88

99.44 ± 0.15

Struct2Graph

99.30 ± 0.22

99.17 ± 0.44

98.19 ± 1.09

99.71 ± 0.13

Method

MCC

F1-score

ROC-AUC

NPV

DeepPPI

95.15 ± 1.55

96.28 ± 1.24

99.27 ± 0.14

98.03 ± 0.78

DeepFE-PPI

97.21 ± 0.72

97.90 ± 0.55

99.51 ± 0.11

99.18 ± 0.29

Struct2Graph

98.20 ± 0.56

98.67 ± 0.41

99.49 ± 0.21

99.33 ± 0.38

  1. Bold face numbers indicate the best performance