benedekrozemberczki / musae Goto Github PK

Dear authors,

As Wikipedia is an open site, is it possible to share the mapping ID2Words of these datasets?

Sincerely,
Bests

Hi, thank you for your great work!
I have further questions.

1. FacebookPagePage dataset

   • I want to reduce its dimension from 128 to 64.
   • So can I get the raw text which you used?
   • I saw your recommendation on issue3. Can I do dimensionality reduction on this dataset, too?

2. Twitch datasets

   • I want to reduce these too.
   • The paper mentions, "Node features are games liked, location and streaming habits."
   • So I think simple dimensionality reduction on this dataset might be harmful.
   • How can I handle these?

Thanks,

Hi Benedek,

I have one question about the file "DE_target.csv". There are several files like this one in the repository.

There are several columns in this file, including "id", "days", "mature", "view", "partner", and "new_id". I am curious about which column indicates the label of a node, that is, whether a streamer uses explicit language.

Could you give me a hint about this? Many thanks!

Best regards,
Simon

In your feature json files, I find that different ids may have different dimensions, why?
If so, how to deal with these features.

In my humble opinion, the matrix corresponding to the undirected graph is symmetric. However, I find it is not the case for the GitHub Social Network(http://snap.stanford.edu/data/github-social.html).
I try to visualize it as follows.

I download the datasets (github) from SNAP, but I'm now confused about the features in .json format.
Have they been preprocessed already so that they can be put into use without further processing?
Or do I need to understand what each dimension in the features mean?

Hi, thanks for your contributions!

About the Facebook dataset - what do the node features represent, and how were they generated? The paper mentions that the features are extracted from site descriptions. Does this mean they're text features, and if so which text representation or embedding did you use?

Thank you for your excellent work! And I would be very grateful if you could answer my question. That is, what's the meaning of the numbers in the node feature json file. For example, in the MUSAE/input/features/git.json. I guess that one vector in the json corresponds to a node, and you mentioned in the manuscript that ` Node features are location, starred repositories, employer and e-mail address'. How can I turn these infomation into the numbers in the json file?

Thank you!