Hyperedge prediction on wordnet non-k-uniform (test)
0.89AUCHyper-SAGNN-E
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Hyper-SAGNN-Etraining_data=hyperedges only2019.11 | 0.89 | 0.705 | |
| Hyper-SAGNN-W (mix)training_data=mixture of edges and hyper-edges2019.11 | 0.884 | 0.684 | |
| Hyper-SAGNN-E (mix)training_data=mixture of edges and hyper-edges2019.11 | 0.881 | 0.694 | |
| Hyper-SAGNN-Wtraining_data=hyperedges only2019.11 | 0.88 | 0.706 | |
| DHNE2019.11 | 0.816 | 0.459 | |
| node2vec - meanAggregation strategy=mean2019.11 | 0.608 | 0.213 | |
| node2vec minAggregation strategy=min2019.11 | 0.575 | 0.2 |