Dataset

[giuliacassara]

Hi @twistedcubic, this is a great code and nice paper!

Can you please provide also cora, dblp and pubmed datasets used in your experiments? Thank you.

[twistedcubic]

Hi @giuliacassara, happy to help.
We originally did not include these data files, as they were too large to host on Github. I found the script we used for converting the DBLP, Cora, and PubMed data from the HyperGCN repo (in particular the pickle files here https://github.com/malllabiisc/HyperGCN/tree/master/data/) (under coauthorship and cocitation) into data files, e.g. dblp_data.pt. We did not include this script in this repo, as this is specific to the HyperGCN data format, and the HNHN code can take in any data file that satisfy the same format as the sample data.

In particular, the create_data function is the data format converter. The script includes how this function is used.

Then the gen_data_dblp() function takes the data files, e.g. dblp_data.pt, unpacks its content to be ready for the HNHN model consumption.