Re: [igraph] Communities in directed networks?

[Tamas Nepusz]

Hi Uri,

> Does igraph has an implemented algorithm for community detection that makes 
> real use of directed informations?
Well, in general it is very hard to come up with a meaningful definition of 
communities in directed networks. Recently, Kim et al have tried to reformulate 
the existing modularity measure in terms of random walks which seems to be more 
suitable for directed networks:

http://arxiv.org/abs/0902.3728

A similar paper from Lambiotte et al with a reformulation in terms of Laplacian 
dynamics, with similar conclusions:

http://arxiv.org/abs/0812.1770

Code to optimize the generalised modularities in this paper are to be found 
here:

http://www.lambiotte.be/codes.html

> If there's no such thing in igraph, where would you recommend looking?
Try the Markov clustering: http://www.micans.org/mcl

Chris Wj has already written some glue code between igraph and MCL using 
igraph's Python interface:
http://igraph.wikidot.com/community-detection-in-python

-- 
Tamas