Node Classification On Wisconsin 60 20 20

評価指標

1:1 Accuracy

評価結果

このベンチマークにおける各モデルのパフォーマンス結果

モデル名	1:1 Accuracy	Paper Title
FAGCN	89.75 ± 6.37	Beyond Low-frequency Information in Graph Convolutional Networks
ACM-Snowball-3	96.62 ± 1.86	Revisiting Heterophily For Graph Neural Networks
APPNP	92.00 ± 3.59	Predict then Propagate: Graph Neural Networks meet Personalized PageRank
GCNII*	89.12 ± 3.06	Simple and Deep Graph Convolutional Networks
ACMII-Snowball-3	97.00 ± 2.63	Revisiting Heterophily For Graph Neural Networks
GCN	75.5 ± 2.92	Semi-Supervised Classification with Graph Convolutional Networks
ACM-GCN+	96.5 ± 2.08	Revisiting Heterophily For Graph Neural Networks
Snowball-3	69.5 ± 5.01	Break the Ceiling: Stronger Multi-scale Deep Graph Convolutional Networks
HH-GCN	79.8 ± 4.30	Half-Hop: A graph upsampling approach for slowing down message passing
Geom-GCN*	64.12	Geom-GCN: Geometric Graph Convolutional Networks
ACM-GCNII*	94.37 ± 2.81	Revisiting Heterophily For Graph Neural Networks
ACMII-Snowball-2	96.63 ± 2.24	Revisiting Heterophily For Graph Neural Networks
ACM-GCNII	94.63 ± 2.96	Revisiting Heterophily For Graph Neural Networks
ACMII-GCN+	96.75 ± 1.79	Revisiting Heterophily For Graph Neural Networks
ACM-Snowball-2	96.38 ± 2.59	Revisiting Heterophily For Graph Neural Networks
SGC-2	74.75 ± 2.89	Simplifying Graph Convolutional Networks
HH-GraphSAGE	85.88 ± 3.99	Half-Hop: A graph upsampling approach for slowing down message passing
HH-GAT	83.53 ± 3.84	Half-Hop: A graph upsampling approach for slowing down message passing
H2GCN	87.5 ± 1.77	Beyond Homophily in Graph Neural Networks: Current Limitations and Effective Designs
GraphSAGE	64.85 ± 5.14	Inductive Representation Learning on Large Graphs

0 of 35 row(s) selected.