Table 2.
Performances (Pearson correlation) on testing set of TT-WOPT and DeepCE with its simpler variants trained with different training sets
| Training sets | Models | Performances |
|---|---|---|
| TT-WOPT | 0.0133 | |
| DeepCE w/o interaction component | 0.4418 | |
| High-quality | DeepCE w/o chemical substructure-gene attention | 0.4620 |
| DeepCE w/o gene-gene attention | 0.4477 | |
| DeepCE | 0.4907 | |
| Augmented | DeepCE | 0.5014 |