Closed nicolaspayette closed 8 years ago
Actually, I just found out that Jung assumes that the graph is strongly connected. I currently have no way of finding out if that is the case. I think that makes it a bug.
Just to be clear, the primitive should work:
Commits 3c11aac and 19e67ef at least turns this issue from a bug to an enhancement request...
Yes, and now it does. See #168
Egeinvector centrality is theoretically not defined for for disconnect networks. It could, however, calculate it separately for each subcomponents.
This is how
closeness centrality
already work, and it would make sense to do the same for egeinvector, it seems.Would be worth seeing what other package do with that (or ask experts).
(By the way, Jung lets you call it on disconnected networks and just gives you nonsensical results.)