Node Classification 벤치마크
Node Classification on genius
Accuracy
- 2016-09-09 — GCN: Accuracy 87.42
- 2019-04-30 — MixHop: Accuracy 90.58
- 2021-10-27 — LINKX: Accuracy 90.77
- 2022-05-15 — GloGNN++: Accuracy 90.91
- 2022-10-14 — ACM-GCN++: Accuracy 91.37
- 2023-01-25 — Dual-Net GNN: Accuracy 91.45
| Rank | Model | Accuracy | 1:1 Accuracy | Paper | Code | Year |
|---|---|---|---|---|---|---|
| 21 | LINK | 73.56 ± 0.14 | – | New Benchmarks for Learning on Non-Homophilous Graphs | CUAI/Non-Homophily-Benchmarks | 2021 |
| 22 | L Prop 2-hop | 67.04 ± 0.20 | – | New Benchmarks for Learning on Non-Homophilous Graphs | CUAI/Non-Homophily-Benchmarks | 2021 |
| 23 | L Prop 1-hop | 66.02 ± 0.16 | – | New Benchmarks for Learning on Non-Homophilous Graphs | CUAI/Non-Homophily-Benchmarks | 2021 |
| 24 | GATJK | 56.70 ± 2.07 | – | New Benchmarks for Learning on Non-Homophilous Graphs | CUAI/Non-Homophily-Benchmarks | 2021 |
| 25 | GAT | 55.80 ± 0.87 | – | Graph Attention Networks | labmlai/annotated_deep_learning_paper_implementations · dmlc/dgl · dmlc/dgl · +90 | 2017 |
| 26 | GESN | – | 91.72 ± 0.08 | Addressing Heterophily in Node Classification with Graph Echo State Networks | dtortorella/addressing-heterophily-gesn | 2023 |