Fig. 6: Comparisons of hypergraph learning efficiency on different hypergraph datasets.

The average training time (s) of 100 epochs of a HGNN, b HGNN+, c HNHN, d HyperGCN, e UniGCN, f UniGAT.

The average training time (s) of 100 epochs of a HGNN, b HGNN+, c HNHN, d HyperGCN, e UniGCN, f UniGAT.