Re: [igraph] Directed Leading Eigenvector

[Chris Wj]

I vote for a feature freeze and alpha release of the 0.6 branch. I think there are some key things implemented in the development branch that are working quite well and are major additions. It's worthy of a release and subsequent releases as new features are added. Frequent releases are healthy for open source projects in general.

-Chris

On Mon, Jul 27, 2009 at 4:53 PM, Tamas Nepusz <address@hidden> wrote:

Is there a specific reason why the directed version of leading
eigenvector hasn't been included?

I think there isn't... as I recall, Gabor ported the symmetric eigensolver first from ARPACK and it wasn't suitable for directed graphs. Now we have an interface to ARPACK's symmetric and nonsymmetric solvers as well (see igraph_arpack_rssolve() and igraph_arpack_rnsolve() in src/arpack.c), so the directed version could be done as well -- we just simply didn't have the time for it. I think it's just a matter of using igraph_arpack_rnsolve, using two degree vectors instead of one (one for the out-degrees and another one for the in-degrees) and adapting the vector-matrix multiplication routine that's used internally by igraph_i_community_leading_eigenvector in src/community.c.

I'd be happy to help write it if need be.

Well, if you are willing to work on it, let me know, I can give you a few pointers to start with and I'll be happy to help should you get stuck. Otherwise I will add the directed version to our ever growing wishlist and try to implement it for the next major release, but I can't tell you when that's gonna happen.

--
Tamas

_______________________________________________
igraph-help mailing list
address@hidden
http://lists.nongnu.org/mailman/listinfo/igraph-help