Closed gaborcsardi closed 11 years ago
indeed, it is relatively easy to write a double loop over clusters, including dealing with different sized groups. it would be even easier if one could specify the parameters and igraph would do it. in particular, there are a few ways of specifying SBM 'parameters', eg,
1) pi, B: B is KxK matrix of elements in (0,1) and pi is a K-dimensional probability vector 2) nvec, B: nvec K-dimensional integer valued vector specifying the # of vertices per cluster
and, of course, one must specify whether it is directed and/or contains self-loops. if self-loops are forbidden, the diagonal blocks must be handled differently than the off-diagonal, and in either case, if things are undirected, the diagonal and off-diagonal must be treated differently.
anyway, you are clearly correct that none of this is difficult, although i think all of it would be useful.
Yes, this issue is about having the algorithm implemented in igraph. :) "Just do an ER..." is for myself, not for you. :)
ah, cool, got it. yes, just do an ER :)
On Tue, Oct 22, 2013 at 9:58 AM, Gabor Csardi notifications@github.comwrote:
Yes, this issue is about having the algorithm implemented in igraph. :) "Just do an ER..." is for myself, not for you. :)
— Reply to this email directly or view it on GitHubhttps://github.com/gaborcsardi/igraph/issues/2#issuecomment-26804780 .
perhaps consider allowing the quest for eudaimonia to guide you openconnecto.me, jovo.me
so it's done?! in master, or some other branch? i'd love to play....
oh, i see now, it is in the development branch, right?
Just do an ER for each pair of blocks.