Node Classification On Penn94

Metrics

Accuracy

Results

Performance results of various models on this benchmark

Model Name	Accuracy	Paper Title	Repository
GPRGCN	81.38 ± 0.16	Adaptive Universal Generalized PageRank Graph Neural Network	-
GATJK	80.69 ± 0.36	New Benchmarks for Learning on Non-Homophilous Graphs	-
GCNII	82.92 ± 0.59	Simple and Deep Graph Convolutional Networks	-
L Prop 2-hop	74.13 ± 0.46	New Benchmarks for Learning on Non-Homophilous Graphs	-
GNNMoE(GCN-like P)	85.11±0.39	Mixture of Experts Meets Decoupled Message Passing: Towards General and Adaptive Node Classification	-
DJ-GNN	84.84±0.34	Diffusion-Jump GNNs: Homophiliation via Learnable Metric Filters	-
C&S 2-hop	78.40 ± 3.12	Combining Label Propagation and Simple Models Out-performs Graph Neural Networks	-
GCNJK	81.63 ± 0.54	New Benchmarks for Learning on Non-Homophilous Graphs	-
C&S 1-hop	74.28 ± 1.19	Combining Label Propagation and Simple Models Out-performs Graph Neural Networks	-
L Prop 1-hop	63.21 ± 0.39	New Benchmarks for Learning on Non-Homophilous Graphs	-
SGC 2-hop	76.09 ± 0.45	Simplifying Graph Convolutional Networks	-
GloGNN	85.57 ± 0.35	Finding Global Homophily in Graph Neural Networks When Meeting Heterophily	-
SGC 1-hop	66.79 ± 0.27	Simplifying Graph Convolutional Networks	-
LINK	80.79 ± 0.49	New Benchmarks for Learning on Non-Homophilous Graphs	-
MixHop	83.47 ± 0.71	MixHop: Higher-Order Graph Convolutional Architectures via Sparsified Neighborhood Mixing	-
H2GCN	81.31 ± 0.60	Beyond Homophily in Graph Neural Networks: Current Limitations and Effective Designs	-
MLP	73.61 ± 0.40	New Benchmarks for Learning on Non-Homophilous Graphs	-
NCGCN	84.74 ± 0.28	Clarify Confused Nodes via Separated Learning	-
ACM-GCN+	85.05 ± 0.19	Revisiting Heterophily For Graph Neural Networks	-
ACMII-GCN+	84.95 ± 0.43	Revisiting Heterophily For Graph Neural Networks	-

0 of 32 row(s) selected.