Discrepancy between processed UCI data provided and referenced UCI Forum data (http://konect.cc/networks/opsahl-ucforum/)

[susheels]

Nice work! I had some questions regarding the datasets used for the experiments.

The UCI dataset mentioned in the paper (Appendix C page 16) references (http://konect.cc/networks/opsahl-ucforum/). The dataset is between users and forums (a bipartite network.) Has 1,421 total nodes with 899 users and 522 forums. It has 33,720 interactions.

However, the processed data provided in this repo and probably used to report numbers in the paper has 1,899 total nodes and 59,835 interactions.

On closer look at the processed dataset I find that you have actually used the UCI Messages dataset (http://konect.cc/networks/opsahl-ucsocial/) dataset which is not bipartite and the interactions are messages between user nodes.

If not, it would be great if you can clarify this discrepancy. Either the description in paper about the UCI dataset is incorrect or the processed dataset provided is wrong.