Node Classification on Cora standard (test)
86AccuracyDFNet-ATT
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| DFNet-ATT2019.10 | 86 | — | — | |
| GCNIItraining time=2m2021.11 | 85.5 | — | — | |
| DFNet2019.10 | 85.2 | — | — | |
| Graph U-NetSource=Reported in literature2019.09 | 84.4 | — | — | |
| CGNNType=Non-linear2021.02 | 84.2 | — | — | |
| GraphMAELearning Setup=unsupervised representation learning with linear classifier2023.06 | 84.2 | — | — | |
| CCA-SSGLearning Setup=unsupervised representation learning with linear classifier2023.06 | 84 | — | — | |
| GraphMix (GCN)Source=Our Experiments, Training Strategy=GraphMix, Backbone=GCN2019.09 | 83.94 | — | — | |
| GATarchitecture=DenseBlock2019.10 | 83.8 | — | — | |
| GCDEType=Non-linear2021.02 | 83.8 | — | — | |
| DisenGCNSource=Reported in literature2019.09 | 83.7 | — | — | |
| GraphHeatType=Non-linear2021.02 | 83.7 | — | — | |
| BVATSource=Reported in literature2019.09 | 83.6 | — | — | |
| MVGRLLearning Setup=unsupervised representation learning with linear classifier2023.06 | 83.5 | — | — | |
| InfoGCLLearning Setup=unsupervised representation learning with linear classifier2023.06 | 83.5 | — | — | |
| ARMA2019.10 | 83.4 | — | — | |
| DF-ATT2019.10 | 83.4 | — | — | |
| DDMLearning Setup=unsupervised representation learning with linear classifier2023.06 | 83.4 | — | — | |
| GraphMix (GAT)Source=Our Experiments, Training Strategy=GraphMix, Backbone=GAT2019.09 | 83.32 | — | — | |
| GraphScanSource=Reported in literature2019.09 | 83.3 | — | — | |
| APPNPType=Non-linear2021.02 | 83.3 | — | — | |
| DGCType=Linear2021.02 | 83.3 | — | — | |
| GATtraining time=1m2021.11 | 83.2 | — | — | |
| GraphSGANCategory=+Adversarial2019.02 | 83 | — | — | |
| GAT2019.10 | 83 | — | — | |
| GATSource=Reported in literature2019.09 | 83 | — | — | |
| GATType=Non-linear2021.02 | 83 | — | — | |
| GATLearning Setup=unsupervised representation learning with linear classifier2023.06 | 83 | — | — | |
| GAT (self-training)Source=Our Experiments, Training Strategy=Self-training2019.09 | 82.95 | — | — | |
| GWWNType=Non-linear2021.02 | 82.8 | — | — | |
| GCNarchitecture=DenseBlock2019.10 | 82.7 | — | — | |
| GATSource=Our Experiments, Training Strategy=Supervised2019.09 | 82.7 | — | — | |
| BGRLLearning Setup=unsupervised representation learning with linear classifier2023.06 | 82.7 | — | — | |
| GraphVATCategory=+Adversarial2019.02 | 82.6 | — | — | |
| iSVD100 (M(NC)) + dropouttraining time=0.1s2021.11 | 82.5 | — | — | |
| DGILearning Setup=unsupervised representation learning with linear classifier2023.06 | 82.3 | — | — | |
| GraphSAGEType=Non-linear2021.02 | 82.2 | — | — | |
| GraphMix (Graph U-Net)Source=Our Experiments, Training Strategy=GraphMix, Backbone=Graph U-Net2019.09 | 82.18 | — | — | |
| Lipschitz normalizationLayers=2, Base Initialization=Xavier2023.10 | 82.1 | — | — | |
| GCN (self-training)Source=Our Experiments, Training Strategy=Self-training2019.09 | 82.03 | — | — | |
| Graph-U-Net (self-training)Source=Our Experiments, Training Strategy=Self-training2019.09 | 82.01 | — | — | |
| iSVD100 (M(NC))training time=0.1s2021.11 | 82 | — | — | |
| CayleyNetdata splitting=as in [21]2019.10 | 81.9 | — | — | |
| MixHoptraining time=26s2021.11 | 81.9 | — | — | |
| GraphSB2026.01 | 81.9 | 81.78 | — | |
| Graph U-NetSource=Our Experiments, Training Strategy=Supervised2019.09 | 81.74 | — | — | |
| MoNetSource=Reported in literature2019.09 | 81.7 | — | — | |
| SIGN-linearType=Linear2021.02 | 81.7 | — | — | |
| GCNscenario=task-specific2018.11 | 81.5 | — | — | |
| GCN2018.02 | 81.5 | — | — | |
| GCN2019.10 | 81.5 | — | — | |
| GCN2019.08 | 81.5 | — | — | |
| GCNSource=Reported in literature2019.09 | 81.5 | — | — | |
| GCNType=Non-linear2021.02 | 81.5 | — | — | |
| GCNtraining time=4s2021.11 | 81.5 | — | — | |
| GCNCategory=+Node Features2019.02 | 81.4 | — | — | |
| GCNSource=Our Experiments, Training Strategy=Supervised2019.09 | 81.3 | — | — | |
| Chebyshev2019.10 | 81.2 | — | — | |
| ChebyshevSource=Reported in literature2019.09 | 81.2 | — | — | |
| JKNetType=Non-linear2021.02 | 81.1 | — | — | |
| SGC-PairNormType=Linear2021.02 | 81.1 | — | — | |
| PSNR-GCNFeature Setting=Fully Observed2023.05 | 81.01 | — | — | |
| SGCType=Linear2021.02 | 81 | — | — | |
| IceBerg2026.01 | 80.99 | 80.95 | — | |
| Half-Hop-GCNFeature Setting=Fully Observed2023.05 | 80.74 | — | — | |
| PSNR-GATFeature Setting=Fully Observed2023.05 | 80.47 | — | — | |
| AdaLNet2019.10 | 80.4 | — | — | |
| DropMessage-GATFeature Setting=Fully Observed2023.05 | 80.36 | — | — | |
| FCN (self-training)Source=Our Experiments, Training Strategy=Self-training2019.09 | 80.3 | — | — | |
| Balanced LL-orthogonal initializationLayers=52023.10 | 80.2 | — | — | |
| ResGCNFeature Setting=Fully Observed2023.05 | 80.11 | — | — | |
| GPT-GNNLearning Setup=unsupervised representation learning with linear classifier2023.06 | 80.1 | — | — | |
| JK-GATFeature Setting=Fully Observed2023.05 | 80.04 | — | — | |
| DropMessage-GCNFeature Setting=Fully Observed2023.05 | 80 | — | — | |
| Res-GATFeature Setting=Fully Observed2023.05 | 79.72 | — | — | |
| Init-Res-GATFeature Setting=Fully Observed2023.05 | 79.67 | — | — | |
| Balanced LL-orthogonal initializationLayers=102023.10 | 79.6 | — | — | |
| DeProp-GCNFeature Setting=Fully Observed2023.05 | 79.52 | — | — | |
| LNet2019.10 | 79.5 | — | — | |
| Balanced LL-orthogonal initializationLayers=22023.10 | 79.5 | — | — | |
| JKNetFeature Setting=Fully Observed2023.05 | 79.42 | — | — | |
| GATFeature Setting=Fully Observed2023.05 | 79.21 | — | — | |
| MTGAEscenario=MTL (Multi-Task Learning)2018.11 | 79 | — | — | |
| GCNFeature Setting=Fully Observed2023.05 | 78.56 | — | — | |
| GraphSHA2026.01 | 78.35 | 77.78 | — | |
| αLoNGAE2018.02 | 78.3 | — | — | |
| CDCGAN2026.01 | 78.12 | 77.92 | — | |
| Lipschitz normalizationLayers=10, Base Initialization=Xavier2023.10 | 78 | — | — | |
| GraphMixup2026.01 | 77.52 | 77.42 | — | |
| GCNIIFeature Setting=Fully Observed2023.05 | 77.4 | — | — | |
| Balanced LL-orthogonal initializationLayers=202023.10 | 77.3 | — | — | |
| Half-Hop-GATFeature Setting=Fully Observed2023.05 | 77.24 | — | — | |
| Lipschitz normalizationLayers=5, Base Initialization=Xavier2023.10 | 77.1 | — | — | |
| DeProp-GATFeature Setting=Fully Observed2023.05 | 76 | — | — | |
| Balanced LL-orthogonal initializationLayers=402023.10 | 75.9 | — | — | |
| GraphENS2026.01 | 75.82 | 74.13 | — | |
| Planetoidscenario=task-specific2018.11 | 75.7 | — | — | |
| Planetoid2018.02 | 75.7 | — | — | |
| PlanetoidCategory=+Node Features2019.02 | 75.7 | — | — | |
| PlanetoidSource=Reported in literature2019.09 | 75.7 | — | — |