[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [igraph] building R package from development branch igraph--main--0.

From: Gabor Csardi
Subject: Re: [igraph] building R package from development branch igraph--main--0.4
Date: Fri, 20 Apr 2007 11:54:13 +0200
User-agent: Mutt/1.5.12-2006-07-14

On Thu, Apr 19, 2007 at 01:54:14PM -0400, Vincent Matossian wrote:
>     Gabor, Tamas,
>    Thanks for the quick response. Gabor, the bootstrap.sh->configure...
>    sequence is exactly what I was looking for. I did encounter problems on
>    cygwin so I will do these builds on linux when needed, but in the meantime
>    use the package you posted, thanks!

I might set up a small script which builds the development version 
every day and uploads it into the webpage. It's easy but there was 
no demand for it yet.
>    I will also look into the community detection algorithms, in your opinion
>    which one would be best to use for a 300k+ node and 1M+ edges graph?

Hmmm, there is a comparison in the walktrap paper mentioned (and written)
earlier by Matthieu Latapy. My guess would be that both the method by 
Clauset, Newman & Moore and maybe also walktrap is able to do this.
Maybe also the leading eigenvector method, especially if you don't 
want all the community hierarchy just a couple of steps. Unfortunately
you cannot set the number of steps to take for calculating the leading 
eigenvector numerically, it is hardcoded in the source right now, 
you might need to play with this if you have large networks.

You can also check these reviews, they might mention something about 
running time:
M. E. J. Newman, Detecting community structure in net-
works. Eur. Phys. J. B 38, (2004). 
L. Danon, J. Duch, A. Diaz-Guilera, and A. Arenas,
Comparing community structure identification. J. Stat.
Mech. p. P09008 (2005).

Please write me if you have some experiences, good or bad, with
the currently implemented methods.

>    I had also been looking for a GML to GraphML converter, is there some code
>    out there that fills that function?

I'm not aware of any, but GML is a very difficult format AFAIK, so 
you can just try to create an edge list from it by standard unix
text processing tools if you're familiar with them.


>    Cheers,
>    Vincent

Csardi Gabor <address@hidden>    MTA RMKI, ELTE TTK

reply via email to

[Prev in Thread] Current Thread [Next in Thread]