Table 5 Performance of AUC values based on multi-task models
From: Large-scale comparison of machine learning methods for profiling prediction of kinase inhibitors
Single models | AUROC | Multi-task models | AUROC |
|---|---|---|---|
GCN | 0.729 | Multi-GCN | 0.785 |
GAT | 0.675 | Multi-GAT | 0.713 |
FP-GNN | 0.704 | FP-GNN | 0.807 |
Chemprop | 0.717 | Chemprop | 0.798 |
AttentiveFP | 0.674 | AttentiveFP | 0.667 |
MPNN | 0.658 | Multi-DNN::Morgan | 0.556 |