Some extreme graphs are hard for the eigensolver. E.g. if you have multiple components, you might as well, just decompose the graph and run community detection on the components.
I suppose we would need a reproducible example to say more. Gabor On Fri, Jan 27, 2017 at 12:33 PM, Karsten Wuellems <[email protected]> wrote: > Hello, > > I use the "community_leading_eigenvector" method and for some graphs I > got some unexpected results. For one community there is a pair of two > vertices that is not connected with any edge to the remaining vertices > of the community. They are just connected via one or two vertices of > another community. Is this possible or is something wrong with my > graph structure? > > Thanks, Karsten > > _______________________________________________ > igraph-help mailing list > [email protected] > https://lists.nongnu.org/mailman/listinfo/igraph-help >
_______________________________________________ igraph-help mailing list [email protected] https://lists.nongnu.org/mailman/listinfo/igraph-help
