Node Classification On Chameleon

Metrics

Accuracy

Results

Performance results of various models on this benchmark

		Paper Title	Repository
DJ-GNN	80.48±1.46	Diffusion-Jump GNNs: Homophiliation via Learnable Metric Filters
FaberNet	80.33±1.19	HoloNets: Spectral Convolutions do extend to Directed Graphs
Dir-GNN	79.71±1.26	Edge Directionality Improves Learning on Heterophilic Graphs
CoED	79.69±1.35	Improving Graph Neural Networks by Learning Continuous Edge Directions
FSGNN (8-hop)	78.27±1.28	Improving Graph Neural Networks with Simple Architecture Design
FSGNN (3-hop)	78.14±1.25	Improving Graph Neural Networks with Simple Architecture Design
HLP Concat	77.48±0.80	Simple Truncated SVD based Model for Node Classification on Heterophilic Graphs	-
Graph ESN	76.2±1.2	Beyond Homophily with Graph Echo State Networks	-
SADE-GCN	75.57±1.57	Self-attention Dual Embedding for Graphs with Heterophily	-
M2M-GNN	75.20 ± 2.3	Sign is Not a Remedy: Multiset-to-Multiset Message Passing for Learning on Heterophilic Graphs
RDGNN-I	74.79 ± 2.14	Graph Neural Reaction Diffusion Models	-
ACMII-GCN++	74.76 ± 2.2	Revisiting Heterophily For Graph Neural Networks
GCNII+DHGR	74.57±2.56	Make Heterophily Graphs Better Fit GNN: A Graph Rewiring Approach	-
ACMII-GCN+	74.56 ± 2.08	Revisiting Heterophily For Graph Neural Networks
UDGNN (GCN)	74.53±1.19	Universal Deep GNNs: Rethinking Residual Connection in GNNs from a Path Decomposition Perspective for Preventing the Over-smoothing	-
ACM-GCN+	74.47 ± 1.84	Revisiting Heterophily For Graph Neural Networks
ACM-GCN++	74.41 ± 1.49	Revisiting Heterophily For Graph Neural Networks
LW-GCN	74.4±1.4	Label-Wise Graph Convolutional Network for Heterophilic Graphs
SignGT	74.31±1.24	SignGT: Signed Attention-based Graph Transformer for Graph Representation Learning	-
CNMPGNN	73.29±1.29	CN-Motifs Perceptive Graph Neural Networks	-

0 of 61 row(s) selected.

Command Palette

Node Classification On Chameleon

Metrics

Results