Node Classification on Cora-ML standard (test)
80Accuracy (Clean)GPRGNN
Evaluation results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| GPRGNNAdv. Training=true, trn. attack=PR-BCD2023.06 | 80 | 79.6 | 79.1 | 79 | 78.7 | 79.8 | 79.3 | 79.3 | 79.1 | |
| GPRGNNAdv. Training=true, trn. attack=LR-BCD2023.06 | 78.8 | 76.5 | 74.5 | 74.1 | 72.5 | 77.7 | 76.2 | 75.3 | 74.5 | |
| ChebNetIIAdv. Training=true, trn. attack=LR-BCD2023.06 | 77.2 | 75 | 71.7 | 71.6 | 69.5 | 75.6 | 74.3 | 73 | 71.7 | |
| ChebNetIIAdv. Training=true, trn. attack=PR-BCD2023.06 | 74.3 | 73.4 | 71.8 | 71.8 | 70 | 73.8 | 72.9 | 72.5 | 71.8 | |
| SoftMedianAdv. Training=false, trn. attack=none2023.06 | 73.2 | 70.3 | 66.8 | 66.5 | 64.6 | 71.7 | 69.5 | 68.1 | 67 | |
| ChebNetIIAdv. Training=false, trn. attack=none2023.06 | 67.5 | 64.4 | 61.2 | 61.3 | 59.1 | 65.8 | 63.5 | 62.3 | 61.5 | |
| APPNPAdv. Training=true, trn. attack=LR-BCD2023.06 | 66 | 57.9 | 48.7 | 47.8 | 43.2 | 60.8 | 55.7 | 51.5 | 49.1 | |
| APPNPAdv. Training=true, trn. attack=PR-BCD2023.06 | 63.2 | 56.3 | 48.1 | 47.4 | 42.6 | 59.7 | 54.2 | 50.6 | 48.1 | |
| GPRGNNAdv. Training=false, trn. attack=none2023.06 | 63 | 55.6 | 47.7 | 47.4 | 43.1 | 59.1 | 54.2 | 49.9 | 47.9 | |
| APPNPAdv. Training=false, trn. attack=none2023.06 | 54.9 | 49.7 | 44.1 | 42.6 | 38.7 | 51.8 | 48.7 | 46.6 | 44.1 | |
| GCNAdv. Training=true, trn. attack=LR-BCD2023.06 | 49.5 | 31.2 | 11.1 | 9.5 | 3.9 | 38.1 | 26.3 | 17.8 | 11.5 | |
| GCNAdv. Training=true, trn. attack=PR-BCD2023.06 | 39.4 | 24.8 | 12.4 | 11.6 | 7.3 | 29.9 | 21 | 16.3 | 12.9 | |
| GCNAdv. Training=false, trn. attack=none2023.06 | 37.6 | 23.6 | 9.6 | 8.8 | 3.8 | 29.9 | 19.7 | 14.3 | 10.1 | |
| GATAdv. Training=true, trn. attack=LR-BCD2023.06 | 32.1 | 29.2 | 21.3 | 20.4 | 14 | 30.5 | 28 | 25.1 | 21.6 | |
| GATAdv. Training=false, trn. attack=none2023.06 | 25.2 | 20 | 6.8 | 4.6 | 0.5 | 22.3 | 19.3 | 11.6 | 6.9 | |
| GATAdv. Training=true, trn. attack=PR-BCD2023.06 | 24.6 | 20.3 | 16.3 | 16.2 | 13.8 | 21.8 | 18.8 | 16.4 | 16.3 |