Table 4 Results of DeepDTA (CNN model) on KIBA and Davis dataset with character-based and multi-granularity encoding. Especially, the character-based encoding methods is original labelling method in DeepDTA [6]
Encoding Method | CI | MSE | \(r^2_m\) | |
|---|---|---|---|---|
KIBA | Character Encoding | 0.863 (0.002) | 0.194 | 0.673 (0.009) |
Multi-Granularity | 0.875 (0.001) | 0.185 | 0.696 (0.017) | |
Davis | Character Encoding | 0.878 (0.004) | 0.261 | 0.630 (0.017) |
Multi-Granularity | 0.884 (0.005) | 0.250 | 0.655 (0.019) |