Node Classification On Wisconsin

Metrics

Accuracy

Results

Performance results of various models on this benchmark

Model Name	Accuracy	Paper Title	Repository
HDP	88.82 ± 3.40	Heterophilous Distribution Propagation for Graph Neural Networks	-
FAGCN	79.61 ± 1.58	Beyond Low-frequency Information in Graph Convolutional Networks
Gen-NSD	89.21 ± 3.84	Neural Sheaf Diffusion: A Topological Perspective on Heterophily and Oversmoothing in GNNs
H2GCN-RARE (λ=1.0)	90.00±2.97	GraphRARE: Reinforcement Learning Enhanced Graph Neural Network with Relative Entropy	-
ACM-GCN+	88.43 ± 2.39	Revisiting Heterophily For Graph Neural Networks
LHS	88.32±2.3	Refining Latent Homophilic Structures over Heterophilic Graphs for Robust Graph Convolution Networks	-
H2GCN-1	84.31 ± 3.70	Beyond Homophily in Graph Neural Networks: Current Limitations and Effective Designs
GloGNN	87.06±3.53	Finding Global Homophily in Graph Neural Networks When Meeting Heterophily
GloGNN++	88.04±3.22	Finding Global Homophily in Graph Neural Networks When Meeting Heterophily
M2M-GNN	89.01 ± 4.1	Sign is Not a Remedy: Multiset-to-Multiset Message Passing for Learning on Heterophilic Graphs
NLGAT	56.9 ± 7.3	Non-Local Graph Neural Networks
ACM-SGC-2	86.47 ± 3.77	Revisiting Heterophily For Graph Neural Networks
GCNH	-	GCNH: A Simple Method For Representation Learning On Heterophilous Graphs
DJ-GNN	-	Diffusion-Jump GNNs: Homophiliation via Learnable Metric Filters
ACM-SGC-1	86.47 ± 3.77	Revisiting Heterophily For Graph Neural Networks
TE-GCNN	87.45 ± 3.70	Transfer Entropy in Graph Convolutional Neural Networks
Geom-GCN-I	58.24	Geom-GCN: Geometric Graph Convolutional Networks
FSGNN (3-hop)	88.43±3.22	Improving Graph Neural Networks with Simple Architecture Design
H2GCN + UniGAP	87.73 ± 4.8	UniGAP: A Universal and Adaptive Graph Upsampling Approach to Mitigate Over-Smoothing in Node Classification Tasks
LINKX	75.49 ± 5.72	Large Scale Learning on Non-Homophilous Graphs: New Benchmarks and Strong Simple Methods

0 of 63 row(s) selected.