BernNet: Learning Arbitrary Graph Spectral Filters via Bernstein Approximation

Benchmark Model Rank Results
node-classification-on-film-60-20-20-random-splitsBernNet#61:1 Accuracy: 41.79 ± 1.01
node-classification-on-non-homophilic-heterophilic-graphs-on-chameleon-60-20-20-random-splitsBernNet#81:1 Accuracy: 68.29 ± 1.58
node-classification-on-chameleon-60-20-20-random-splitsBernNet#91:1 Accuracy: 68.29 ± 1.58
node-classification-on-non-homophilic-heterophilic-graphs-on-texas-60-20-20-random-splitsBernNet#131:1 Accuracy: 93.12 ± 0.65
node-classification-on-squirrel-60-20-20-random-splitsBernNet#131:1 Accuracy: 51.35 ± 0.73
node-classification-on-texas-60-20-20-random-splitsBernNet#141:1 Accuracy: 93.12 ± 0.65
node-classification-on-cornell-60-20-20-random-splitsBernNet#151:1 Accuracy: 92.13 ± 1.64
node-classification-on-non-homophilic-heterophilic-graphs-on-cornell-60-20-20-random-splitsBernNet#151:1 Accuracy: 92.13 ± 1.64
node-classification-on-cora-60-20-20-random-splitsBernNet#181:1 Accuracy: 88.52 ± 0.95
node-classification-on-citeseer-60-20-20-random-splitsBernNet#211:1 Accuracy: 80.09 ± 0.79
node-classification-on-pubmed-60-20-20-random-splitsBernNet#261:1 Accuracy: 88.48 ± 0.41