[igraph] Re: Suggestion for a new community detection function

[Simone Caschili]

Dear Tamas,

Well you right, the approach they propose are different compared to
Newman et al. one. What looked similar to me is the starting point:
both algorithms assume that  each node is a sole community at the
first step. As far as I have understood, in a interactive way, both
algorithms join communities together choosing among those with higher
modularity value. What changes is how the calculate the modularity.

Well, probably I'm misunderstanding the meaning of the paper
arXiv:cond-mat/0408187v2, that is really technical for my knowledges!

Thank you to take into account my suggestion to implement that
algorithm in the near future!

Regards,
Simon

Quoting Tamas
> Dear Simon,
>
>> I'm starting implementing some analyses through community detection
>> algorithms. I have found interesting a new version of modularity
>> proposed Lancichinetti et al. (2008).
> I read that paper and I'm a little bit confused - seemingly they do
> not suggest a new modularity measure, but a new algorithm to discover
> overlapping communities in complex networks using a local fitness
> measure of individual vertices. (So the fitness measure belongs to a
> given community and a single vertex in relation to that community, and
> vertices are added to or removed from the community iteratively based
> on this fitness measure). I checked the algorithm and I think it is
> not too difficult to implement in Python - if I have some time, I'll
> give it a go and see if I can come up with a simple implementation;
> however, this has nothing to do (IMHO) with the original modularity
> measure of Newman et al and it requires a completely different approach.
>
> --
> T.