search

micans / mcl

MCL, the Markov Cluster algorithm, also known as Markov Clustering, is a method and program for clustering weighted or simple networks, a.k.a. graphs.
https://micans.org/mcl
Other
89 stars 12 forks source link

Two issues: "nx.to_scipy_sparse_matrix" not exist and the hyperparameter "inflation" not working? #16

Closed HsiangHung closed 1 year ago

HsiangHung commented 1 year ago

Hi MCL Community,

I found two issues:

  1. In the sample code, nx.to_scipy_sparse_matrix doesn't exist in networkx 3.0. The document needs to be updated to nx.to_scipy_sparse_array.
  2. I tried the doc: https://markov-clustering.readthedocs.io/en/latest/readme.html and found different hyperparameter inflation =1.5-2.6 also gave the same results and modularity are different from the doc page: image

Also inflation=1.1 and inflation=2.5 gave the same clustering outcome. I think there is a bug. Can you check? Thanks.

micans commented 1 year ago

Hello,

I see you have also opened https://github.com/GuyAllard/markov_clustering/issues/37 - that is the correct place for this issue as it pertains to the code in that repository. This repository is an entirely different code base. I will close the issue here, and will participate in https://github.com/GuyAllard/markov_clustering/issues/37 if that is useful. Please note I have not been involved with that code however, and secondly the code does not seem to be actively supported anymore. If you are not committed to a Python implementation you could give the C implementation here a try, although I understand that this could be inconvenient.

micans commented 1 year ago

(Feel free to ask further questions here, but I cannot help with the issue itself)