I have a 3D surface from which I extract certain spatial parameters like extremas and their locations (a.k.a coordinates), which then I convert into a bipartite graph with distance between them as edges. Now, using MST, I reduce it to a tree.

This end structure is supposed to represent a lower dimensional version of the above surface. Now, I assign a label to the graph (actually, I get two graphs for 1 label)

I need to pass this data set to a supervised classifier to get an output. But, CNNs and RNNs seem not to work on graph data. Is there a way to input Graph and get a classified label?

  • 1
    $\begingroup$ Could you share an example of the initial data? Why do you want to use a graph first to classify? I ask this because in some cases, you can directly classify without going through a graph. $\endgroup$ Aug 3, 2021 at 7:42
  • $\begingroup$ I wanted to make it into a graph because it is really memory intensive. $\endgroup$ Aug 3, 2021 at 10:17
  • $\begingroup$ @NicolasM datascience.stackexchange.com/questions/97555/… $\endgroup$ Aug 3, 2021 at 10:17
  • $\begingroup$ @NicolasM my other question is about that. I flattened it into a 2D image from top view. Which I passed into a CNN with labels, it was really memory intensive and I kept filling my disk because I was using memory mappings and they were overflowing in RAM. $\endgroup$ Aug 3, 2021 at 10:19
  • $\begingroup$ Also, I'll share the data when I get home as I don't have internet access to my workstation in office. $\endgroup$ Aug 3, 2021 at 10:20


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.