use random sampling to compute centrality estimates
Bug #1090728 reported by
Jan Katins
This bug affects 1 person
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
igraph |
Confirmed
|
Wishlist
|
Unassigned |
Bug Description
betweeness centrality is not useable for networks with >100k vertices (a |V|=200k, |E|=1.2mio network and `g.betweenness(
CENTRALITY ESTIMATION IN LARGE NETWORKS
ULRIK BRANDES and CHRISTIAN PICH
International Journal of Bifurcation and Chaos 2007 17:07, 2303-2318
http://
http://
To post a comment you must log in.
Thanks for the suggestion.
I think that you can easily do this "by hand". Just sample a number of vertices, calculate shortest paths individually from the selected vertices, and aggregate the results. I haven't read those papers tough, so I might be missing something.