Table 2.

Performance evaluation of the DTA prediction models on the Davis dataset.

Methods	Proteins	Compounds	MSE ↓	CI ↑	${\mathit{r}}_{\mathit{m}}^2$↑	Pearson ↑
KronRLS	Smith–Waterman	Pubchem-Sim	0.379	0.871	0.407	-
SimBoost	Smith–Waterman	Pubchem-Sim	0.282	0.872	0.644	-
DeepDTA	CNN	CNN	0.261	0.878	0.630	-
WideDTA	CNN + PDM	CNN + LMCS	0.262	0.886	-	0.820
GraphDTA	CNN	GIN	0.229	0.893	-	-
GEFA	GCN	GCN	0.228	0.893	-	0.847
MGraphDTA	MCNN	MGNN	0.207	0.900	0.710	-
WGNN-DTA	GCN	GCN	0.208	0.900	0.692	0.861
WGNN-DTA	GAT	GAT	0.208	0.903	0.691	0.863
DGraphDTA	GCN	GCN	0.202	0.904	0.700	0.867
MSGNN-DTA	GAT	GAT + GAT	0.195	0.906	0.719	0.871

Note: Bold indicates the best result in the evaluation metrics. These results are not reported from original studies.