Re: [igraph] Does igraph random_walk function work with weighted networks?

[Szabolcs Horvát]

On 10 November 2017 at 00:47, Olli Pietilainen <address@hidden> wrote:

Hello,

I'm interested in using  random walk function in igraph R package to study protein-protein interaction network.

I was wondering does the random_walk function handle weighted networks to modify the transition probability from one vertex to another or are all transitions always equally likely to happen?

Currently, this function does not support weights (although I noticed that there was a TODO note in the C source code to add this).

I am mainly commenting to suggest a design improvement, in case weights get added in the future:

This function should return a set of traversed edges, not a set of vertices.  It is a fairly common application to do a random walk on multigraphs (i.e. graphs having parallel edges), and then make some decision based on how frequently edges were traversed.  Different edges between the very same vertices may be assigned different effects.

 

Thank you so much for your help!

Best,

Olli Pietilainen

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