Closed ghost closed 4 years ago
hi, thank you for your interest for SGC.
-- Is is possible? Yes! I believe if you just change your adjacency matrix and follow the code in this repo, it should work.
-- will it work well? I don't really know. I think it is very difficult to know before actually trying it out and personally I havn't tried it. It will probably, just like many other machine learning problems, depend on the dataset you are working with.
But I encourage to try SGC out! If you are starting to work on something new, why not try this simple and fast baseline? Good luck!
GCN can handle directed graph when adjacency matrix is allowed to be asymmetric and normalized with D^-1 * A. Is it possible to have SGC handle directed graph as well? Thanks