Re: [igraph] Compare clusterings (partitions of a set)

[Tamas Nepusz]

Hi all,

> This is something I am quite interested in. There are two metric 
> dissimilarities
> (also called metrics or distances) between clusterings that I think are very
> useful.  One is the 'variation of information' measure by Meila, also provided
> in cl_dissimilarity.
Just for the record: the variation of information metric (Meila et al, 2003) 
and the normalized mutual information measure (Danon et al, 2005) will be 
included in igraph 0.6; they are already available in the Bazaar repository as 
igraph_compare_communities (or simply compare_communities in the Python 
interface, compare.communities in the R interface).

I'll read your technical report about the split/join distance.

-- 
Tamas