The data/ folder contains all the datasets as well as the intermediate hyperedge features needed for the 3 datasets, where the node features are initialized randomly.
The src/ folder contains all the scripts for the different datasets used by our model.
Instructions:
- To reproduce the results for High-school, change the seed in line 37 of run_highschool.py and run the script.
- To reproduce the results of Cora and DBLP, change line 18 of ihnn_cora.py and ihnn_dblp.py and run the scripts.
- To reproduce the results of walmart-trips dataset, change the seed in line 40 of run_walmart.py and run the script.