Table 3.

Performance comparison between our model and existing methods on the random splitting human dataset

Model	AUC	Recall	Precision
k-NNa	0.860	0.927	0.798
RFa	0.940	0.897	0.861
L2a	0.911	0.913	0.861
SVMb	0.910	0.966	0.969
GNNa	0.970	0.918	0.923
GCNb	0.956 ± 0.004	0.862 ± 0.006	0.928 ± 0.010
GraphDTAb	0.960 ± 0.005	0.882 ± 0.040	0.912 ± 0.040
TransformerCPIb	0.973 ± 0.002	0.916 ± 0.006	0.925 ± 0.006
DrugVQAa	0.979 ± 0.003	0.961 ± 0.002	0.954 ± 0.03
MINN-DTI	0.981 ± 0.003	0.945 ± 0.030	0.902 ± 0.045

^aMeans results obtained from the article (Zheng et al., 2020).

^bMeans results obtained from the article (Chen et al., 2020).

Best results of the corresponding experiments were represented in bold.