> Since infomap is random-walk based, I _think_ the code would require some > accommodation for disconnected components (like tunneling or code to > iterate over components). But it's also possible that infomap has been > updated since that comment was written and it's no longer relevant. AFAIK the version of Infomap in igraph is essentially the same as an older implementation written by Martin Rosvall. He has rewritten Infomap since and released the new version under a license which is incompatible with igraph's license, hence we have to stick to the old version in igraph. I don't know what limitations the old version has, but since it's the same as Martin's old implementation, I guess the limitations are also the same.
T. _______________________________________________ igraph-help mailing list [email protected] https://lists.nongnu.org/mailman/listinfo/igraph-help
