CODEWITHSUNDEEP
pythonigraph
skyork
skyork

Reputation: 7391

Community detection and comparison in igraph

I am using the multilevel algorithm in igraph to detect communities among the same set of the nodes over a period of time (i.e. multiple and potentially different clusterings will be produced for the same set of nodes). I have two questions regarding the output of the detection algorithm (which may also apply to other detection algorithms in igraph):

  1. How does the algorithm decide the membership assignment? In other words, is there any rule or pattern that decides which nodes are assigned to the 1st subgraph of the resulted VertexClustering object and which ones will be assigned to the 2nd subgraph, and so on?

  2. If there is some consistent membership assignment pattern, is there a method to compare the membership of two communities that are detected at different times?

Upvotes: 0

Views: 2009

Answers (1)

Gabor Csardi
Gabor Csardi

Reputation: 10825

  1. See the paper referenced in the documentation: http://igraph.sourceforge.net/doc/python/igraph.Graph-class.html#community_multilevel

  2. See http://igraph.sourceforge.net/doc/python/igraph.clustering-module.html#compare_communities

Upvotes: 1

Related Questions