Re: [igraph] Leading eigenvector for directed graphs

[Tadaishi Yatabe-Rodriguez]

Thanks Tamas. Regarding the InfoMap algorithm. Does it include edge direction in spite not being explicitly in the R code?

Thanks,

Tada

On Sat, Aug 30, 2014 at 12:34 PM, Tamás Nepusz <address@hidden> wrote:

Hi,

> Has it been implemented yet this algorithm for directed paths?

No, mostly because there are several possible extensions of the modularity measure (that the leading eigenvector method tries to optimize) for directed graphs, and there is no consensus yet on which possible extension is the "best" and how the leading eigenvector method could be adapted for the "best" directed modularity measure.

> I have a directed weighted network of 62 nodes and 130 arcs, and can't made
> up my mind yet of what to use. The leading eigenvector seems nice, but
> can't find how to actually run it.

If you are looking for an algorithm that uses some kind of a modularity measure, this seems to be a neat one:

http://stat.kaist.ac.kr/papers/2010_LinkRank.pdf
(no implementation in igraph, although the authors might have published one)

Otherwise, I would suggest the excellent InfoMap algorithm (see http://www.mapequation.org), which is not only implemented in igraph but it also has a theoretical foundation that handles directed links naturally.

All the best,
T. 

--

Tadaishi Yatabe

DVM, MPVM, PhD (C)

Center for Animal Disease Modeling and Surveillance (CADMS)

Department of Medicine and Epidemiology

School of Veterinary Medicine

University of California Davis