Node classification on Texas (test)
94.8Mean AccuracyNode2Vec
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| Node2Vec2026.06 | 94.8 | — | — | — | |
| BernNet2023.09 | 93.12 | — | — | — | |
| GPRGNN2023.09 | 92.95 | — | — | — | |
| GPRGNNModel Type=Deep2022.02 | 92.92 | — | — | — | |
| MLPModel Type=Deep2022.02 | 92.26 | — | — | — | |
| HiGCNOrder=22023.09 | 92.15 | — | — | — | |
| HiGCNOrder=32023.09 | 91.85 | — | — | — | |
| MLP2023.09 | 91.45 | — | — | — | |
| HiGCNOrder=42023.09 | 91.42 | — | — | — | |
| APPNPModel Type=Deep2022.02 | 91.18 | — | — | — | |
| APPNP2023.09 | 90.98 | — | — | — | |
| ASPECTEvaluation Protocol=linear protocol2026.04 | 90.9 | — | — | — | |
| TopoU-Net2026.05 | 90.8 | — | — | — | |
| HiGCNOrder=12023.09 | 90.36 | — | — | — | |
| DiagPolySD2025.11 | 90 | — | — | — | |
| SGAT2023.09 | 89.83 | — | — | — | |
| BundlePolySD2025.11 | 89.74 | — | — | — | |
| PPGNNGraph modification variant=Original2022.06 | 89.7 | — | — | — | |
| SGATEF2023.09 | 89.67 | — | — | — | |
| GeneralPolySD2025.11 | 89.21 | — | — | — | |
| POLYGCLEvaluation Protocol=linear protocol2026.04 | 88.03 | — | — | — | |
| RiSNNtransport=none2025.11 | 87.89 | — | — | — | |
| CCA-SSGEvaluation Protocol=linear protocol2026.04 | 87.87 | — | — | — | |
| DeepWalk2026.06 | 87.6 | — | — | — | |
| G2-GraphSAGEgradient_gating=true2022.10 | 87.57 | — | — | — | |
| JdSNN2025.11 | 87.37 | — | — | — | |
| JdSNNweighting=none2025.11 | 87.3 | — | — | — | |
| RiSNN2025.11 | 86.84 | — | — | — | |
| ASGCModel Type=Non-Deep2022.02 | 86.76 | — | — | — | |
| RawModel Type=Non-Deep2022.02 | 86.49 | — | — | — | |
| WG-SRCn=102026.04 | 86.32 | 1.99 | — | 4.08 | |
| FGN2026.06 | 86.2 | — | — | — | |
| Conn-NSDModel Category=Sheaf-based2022.06 | 86.16 | — | — | — | |
| Conn-NSD2025.11 | 86.16 | — | — | — | |
| S3GCLEvaluation Protocol=linear protocol2026.04 | 86.12 | — | — | — | |
| ChebNet2023.09 | 86.08 | — | — | — | |
| GCN-ChebyModel Type=Deep2022.02 | 86.08 | — | — | — | |
| O(d)-NSDModel Category=Sheaf-based2022.06 | 85.95 | — | — | — | |
| HH-GraphSAGEBackbone=GraphSAGE, Half-Hop augmentation=true2023.08 | 85.95 | 3.51 | — | — | |
| Sheaf2024.10 | 85.95 | — | — | — | |
| NSDrestriction maps=orthogonal2025.11 | 85.95 | — | — | — | |
| GGCN2023.09 | 85.81 | — | — | — | |
| ANSD2025.11 | 85.68 | — | — | — | |
| Diag-NSDModel Category=Sheaf-based2022.06 | 85.67 | — | — | — | |
| NSDrestriction maps=diagonal2025.11 | 85.67 | — | — | — | |
| k-S2V2023.09 | 85.12 | — | — | — | |
| H2GCN2026.06 | 85.11 | — | — | — | |
| H2GCNGraph modification variant=Original2022.06 | 84.9 | — | — | — | |
| GCN-IED2026.06 | 84.9 | — | — | — | |
| GGCNModel Category=GNN2022.06 | 84.86 | — | — | — | |
| H2GCNModel Category=GNN2022.06 | 84.86 | — | — | — | |
| H2GCNbest model variant=true2021.06 | 84.86 | — | — | — | |
| H2GCNAttack=PGD, Perturbation=0.00012022.02 | 84.86 | — | — | — | |
| H2GCN-12022.05 | 84.86 | — | — | — | |
| GGCN2022.10 | 84.86 | — | — | — | |
| H2GCN2022.10 | 84.86 | — | — | — | |
| G2-GCNgradient_gating=true2022.10 | 84.86 | — | — | — | |
| GGCNHalf-Hop augmentation=false2023.08 | 84.86 | — | — | — | |
| H2GCNHalf-Hop augmentation=false2023.08 | 84.86 | — | — | — | |
| GGCN2024.10 | 84.86 | — | — | — | |
| GGCN2025.11 | 84.86 | — | — | — | |
| H2GCN2025.11 | 84.86 | — | — | — | |
| CMGNN2026.06 | 84.68 | — | — | — | |
| UDGNNBackbone=GCN2022.05 | 84.6 | — | — | — | |
| GPRGNNGraph modification variant=Original2022.06 | 84.6 | — | — | — | |
| G2-GATgradient_gating=true2022.10 | 84.59 | — | — | — | |
| CoEDedge direction learning=false2024.10 | 84.59 | — | — | — | |
| EIGNNAttack=FGSM, Perturbation=0.00012022.02 | 84.33 | — | — | — | |
| EIGNNAttack=PGD, Perturbation=0.00012022.02 | 84.33 | — | — | — | |
| GloGNN2022.10 | 84.32 | — | — | — | |
| A2DUG2023.06 | 84.32 | — | — | — | |
| GraphSAGEn=102026.04 | 84.32 | — | — | 4.73 | |
| H2GCN+LCC2026.06 | 84.32 | — | — | — | |
| GPRGNNGraph modification variant=Adaptive Spectral Clustering2022.06 | 84.3 | — | — | — | |
| GloGNN2026.06 | 84.3 | — | — | — | |
| RandEdge-NSDModel Category=Sheaf-based2022.06 | 84.05 | — | — | — | |
| UDGNNBackbone=SAGE2022.05 | 84.05 | — | — | — | |
| SAN2025.11 | 84.05 | — | — | — | |
| EIGNNAttack=FGSM, Perturbation=0.0012022.02 | 83.79 | — | — | — | |
| EIGNNAttack=PGD, Perturbation=0.0012022.02 | 83.79 | — | — | — | |
| DirGNN2024.10 | 83.78 | — | — | — | |
| DSHN2025.10 | 83.78 | — | — | 5.13 | |
| WRGAT2021.06 | 83.62 | — | — | — | |
| Co-GNN2024.10 | 83.51 | — | — | — | |
| HyperND2025.10 | 83.51 | — | — | 5.19 | |
| UDGNNBackbone=GAT2022.05 | 83.43 | — | — | — | |
| MagNet2024.10 | 83.3 | — | — | — | |
| FSGNN2023.06 | 83.24 | — | — | — | |
| Magnet2023.06 | 83.24 | — | — | — | |
| DHGNN2025.10 | 83.24 | — | — | 5.64 | |
| BernNetGraph modification variant=Original2022.06 | 83.2 | — | — | — | |
| RandNode-NSDModel Category=Sheaf-based2022.06 | 82.97 | — | — | — | |
| Gen-NSDModel Category=Sheaf-based2022.06 | 82.97 | — | — | — | |
| UDGNN*Backbone=SAGE2022.05 | 82.97 | — | — | — | |
| NSDrestriction maps=general2025.11 | 82.97 | — | — | — | |
| AllDeepSets2025.10 | 82.76 | — | — | 5.74 | |
| AllSetTransformer2025.10 | 82.76 | — | — | 5.07 | |
| F2GAT2022.10 | 82.7 | — | — | — | |
| All featuresselection_ratio=100%, backbone=GNN2025.10 | 82.7 | — | — | — | |
| GGDEvaluation Protocol=linear protocol2026.04 | 82.62 | — | — | — |