Hi Gábor,

> * Gábor Csárdi <[email protected]> [2012-08-14 
> 11:39:07 -0400]:
>
> the spinglass methos has a resolution parameter (called gamma), so
> tuning that should give you different results. Try to increase it from
> the default 1. Hopefully spinglass won't take too long to run on your
> data.

increasing it (even to 100) only gives me 25 communities with a very
uneven distribution of members:

   1    2    3    4    5    6    7    8    9   10   11   12   13   14   15   16 
 269  234  252  241  136   67  135  165  631  139   29   82  250    1  377  318 
  17   18   19   20   21   22   23   24   25 
 575   51   61  115  171  168 1762  518  208 

(for a highly pruned graph with v=6955 e=214234)

I actually have vertex weights in addition to edge weights; is there a
community detection algorithm which would try to make total community
weights approximately the same?

-- 
Sam Steingold (http://sds.podval.org/) on Ubuntu 12.04 (precise) X 11.0.11103000
http://www.childpsy.net/ http://dhimmi.com http://ffii.org http://memri.org
http://camera.org http://www.memritv.org http://pmw.org.il
Do not tell me what to do and I will not tell you where to go.

_______________________________________________
igraph-help mailing list
[email protected]
https://lists.nongnu.org/mailman/listinfo/igraph-help

Reply via email to