Hypergraph Node Classification on Citeseer (test)
66.6AccuracyHyperGCL (A6: proposed generative augmentation)
Evaluation results
| Method | Links | |
|---|---|---|
| HyperGCL (A6: proposed generative augmentation)Attack Type=Random, Training Ratio=10%2022.10 | 66.6 | |
| HyperGCL (A2: generalized hyperedge augmentation)Attack Type=Random, Training Ratio=10%2022.10 | 66.41 | |
| HyperGCL (A4: feature perturbation)Attack Type=Random, Training Ratio=10%2022.10 | 65.94 | |
| HyperGCL (A6: proposed generative augmentation)Attack Type=Net, Training Ratio=10%2022.10 | 65.68 | |
| HyperGCL (A6: proposed generative augmentation)Attack Type=Minmax, Training Ratio=10%2022.10 | 65.51 | |
| HyperGCL (A2: generalized hyperedge augmentation)Attack Type=Net, Training Ratio=10%2022.10 | 65.38 | |
| HyperGCL (A4: feature perturbation)Attack Type=Net, Training Ratio=10%2022.10 | 65.15 | |
| HyperGCL (A2: generalized hyperedge augmentation)Attack Type=Minmax, Training Ratio=10%2022.10 | 64.69 | |
| HyperGCL (A4: feature perturbation)Attack Type=Minmax, Training Ratio=10%2022.10 | 64.12 | |
| SetGNNAttack Type=Random, Training Ratio=10%2022.10 | 62.89 | |
| SetGNNAttack Type=Net, Training Ratio=10%2022.10 | 62.81 | |
| SetGNNAttack Type=Minmax, Training Ratio=10%2022.10 | 62.21 |