Efficient Heterogeneous Graph Learning via Random Projection

Benchmark Model Rank Results
heterogeneous-node-classification-on-acmRpHGNN#1Macro-F1: 94.09Micro-F1: 94.04
heterogeneous-node-classification-on-dblp-2RpHGNN#1Macro-F1: 95.23Micro-F1: 95.55
heterogeneous-node-classification-on-freebaseRpHGNN#1Macro-F1: 54.02Micro-F1: 66.55
heterogeneous-node-classification-on-imdbRpHGNN#1Macro-F1: 67.53Micro-F1: 69.77
heterogeneous-node-classification-on-oagRpHGNN#1NDCG: 53.31MRR: 35.46
heterogeneous-node-classification-on-oag-l1RpHGNN#1NDCG: 87.80MRR: 86.79
node-property-prediction-on-ogbn-magRpHGNN+LP+CR (LINE embs)#3Test Accuracy: 0.5773 ± 0.0012Ext. data: No