Table 3.

MCNN-DDI results on different sets of features.

Set of Features	Similarity Matrices	ACC	F1	AUC	AUPR
S	Jaccard	0.8861	0.8099	0.9983	0.9505
T	Jaccard	0.8441	0.7517	0.9977	0.9254
E	Jaccard	0.6802	0.4584	0.9920	0.7691
P	Jaccard	0.8317	0.7357	0.9975	0.9169
S + T	Jaccard	0.8986	0.8264	0.9985	0.9549
S + E	Jaccard	0.8933	0.7943	0.9976	0.9309
S + P	Jaccard	0.8993	0.8294	0.9986	0.9552
T + P	Jaccard	0.8509	0.7667	0.9980	0.9313
T + E	Jaccard	0.8658	0.7498	0.9972	0.9158
P + E	Jaccard	0.8635	0.7563	0.9973	0.9160
S + P + E	Jaccard	0.8998	0.8102	0.9981	0.9464
S + T + E	Jaccard	0.9000	0.8286	0.9981	0.9478
S + T + P	Jaccard	0.8849	0.8077	0.9986	0.9530
T + P + E	Jaccard	0.8672	0.7833	0.9976	0.9292
S + T + E + P	Jaccard	0.8953	0.8164	0.992	0.9503