Re: [igraph] Assortativity in igraph-6.0 under R

[Gang Su]

My previous email to the list seemed didn't go through, so I will just 
send it again.
---------
Hi Gabor,

The authors used a two-step algorithm which includes 1. recursive 
partition. 2. refinement.
The refinement part is independent of the community algorithm. And 
actually this scheme can be applied to any community algorithm which
can work on disconnected graphs (also we can have a fix to let it work 
on connected graphs).

It gives smaller communities for larger networks.
Actually it's not very difficult to implement this 2-step procedure in 
R, but it won't be very efficient.
So I was wondering actually instead of Q-cut, do you have a plan for 
implementation of such
multi-step procedures. I can be very helpful because it's generic, and 
can actually work on any graph partitioning algorithm.

Gang

On 5/19/2010 10:47 AM, Gábor Csárdi wrote:
> I think what you want to compare is
>
>    
> > assortativity(g1,degree(g1, mode="out")-1, degree(g1, mode="in")-1, 
> > directed=TRUE)
> >      
> [1] -0.3131201
>    
> > assortativity.degree(g1, directed=TRUE)
> >      
> [1] -0.3131201
>
> because this is how assortativity.degree() works if directed=TRUE.
>
> Best,
> Gabor
>
> On Wed, May 19, 2010 at 4:36 PM, Przemek Grabowicz
> <address@hidden>  wrote:
>    
> > Thank you for implementing assortativity in igraph-0.6, it's nice addition.
> >
> > However there is something there what I don't understand.
> > Assortativity.degree gives different results than assortativity itself. If I
> > understand manual properly it should give the same values.
> >
> >      
> > > assortativity(g1,degree(g1)-1,directed=TRUE)
> > >        
> > [1] 0.1368699
> >      
> > > assortativity.degree(g1,directed=TRUE)
> > >        
> > [1] 0.1546312
> >
> > It works fine when directed=FALSE. I'm using
> > igraph_nightly_0.6-1899-20100505 under R-2.8.1 .
> >
> > _______________________________________________
> > igraph-help mailing list
> > address@hidden
> > http://lists.nongnu.org/mailman/listinfo/igraph-help
> >
> >      
>
>
>