Skip to main content

Table 8 Comparative results of SSF-DDI (sequence and substructure features for drug–drug interactions prediction) on the DrugBank dataset (%)

From: SSF-DDI: a deep learning method utilizing drug sequence and substructure features for drug–drug interaction prediction

Method

ACC

AUC

F1

Prec

Rec

AP

SF-DDI w/o substructure

94.77

98.34

94.92

92.23

97.76

97.83

SF-DDI w/o sequence

96.09

98.78

96.15

94.74

97.59

98.29

SSF-DDI w/o SGFE and sequence

95.56

98.58

95.63

94.05

97.27

98.15

SSF-DDI(ours)

96.45

98.92

96.5

95.22

97.89

98.53

  1. The best results are highlighted in bold