Graph Classification on D&D Biochemical TU (10-fold cross val)
78.74AccuracyHyperDrop
Evaluation results
| Method | Links | |
|---|---|---|
| HyperDropArchitecture Type=Ours, Evaluation Protocol=10-fold CV2021.06 | 78.74 | |
| GMTArchitecture Type=Global, Evaluation Protocol=10-fold CV2021.06 | 78.72 | |
| HyperDrop + GMTArchitecture Type=Ours, Evaluation Protocol=10-fold CV2021.06 | 78.39 | |
| MinCutPoolArchitecture Type=Hierarchical, Evaluation Protocol=10-fold CV2021.06 | 78.22 | |
| DiffPoolArchitecture Type=Hierarchical, Evaluation Protocol=10-fold CV2021.06 | 77.56 | |
| DeepSetArchitecture Type=Set, Evaluation Protocol=10-fold CV2021.06 | 77.39 | |
| ASAPArchitecture Type=Hierarchical, Evaluation Protocol=10-fold CV2021.06 | 76.58 | |
| EdgePoolArchitecture Type=Hierarchical, Evaluation Protocol=10-fold CV2021.06 | 75.85 | |
| SortPoolArchitecture Type=Global, Evaluation Protocol=10-fold CV2021.06 | 75.58 | |
| SAGPoolArchitecture Type=Hierarchical, Evaluation Protocol=10-fold CV2021.06 | 74.72 | |
| TopKPoolArchitecture Type=Hierarchical, Evaluation Protocol=10-fold CV2021.06 | 73.63 | |
| GCNArchitecture Type=Naive GNN, Evaluation Protocol=10-fold CV2021.06 | 72.05 | |
| GINArchitecture Type=Naive GNN, Evaluation Protocol=10-fold CV2021.06 | 70.79 |