I'm looking for articles and other reading material on handling and modeling graph-shaped data for classification systems. As I've never working with graphed base machine learning anything from novice to advanced would be good. I'm after solving graph problems, but representing data that natively fits into a graph diagram in a way that would benefit machine learning on it.

I would like to experiment with predicting based on a graph of directed interactions between nodes in a system as well as incidental events occurring on one or more of the nodes (In hopes of deriving information from correlations between # of nodes similar incidents happened at, relations between nodes with similar incidents, etc).

Anything from published articles to experience tips and best practices would be highly appreciated.

  • $\begingroup$ You want to predict a property of the nodes using covariates at the nodes with a correlation structure defined by the graph edges? That's a standard autologistic regression problem. $\endgroup$
    – Spacedman
    Nov 15, 2016 at 23:29

1 Answer 1


Super comprehensive question. You are basically asking for tones of directions! I'd suggest to start with link prediction problem. So assume you have a directed/signed/weighted graph and you want to find the most probable next interaction (link/edge). I'd suggest to have a look at Jerome Kunegis papers and you may see it as a starting point and go on from there.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.