Re: [igraph] Working with large networks and how to sample from a graph?

[Peter Flom]

Thanks Tamas

That brings up a suggestion - in the next version of the manual it would be
nice to see, for each measure, a rough idea of running time. Obviously, this
couldn't be precise as it depends on the machine running the code and
probably on lots of other things, but something indicating that betweenness
centrality running time is proportional to links*nodes, whereas degree seems
to be linear in nodes (from what I can tell) and similar for other measures.

Thanks again

Peter Flom
Peter Flom Consulting
http://www.statisticalanalysisconsulting.com/
http://www.IAmLearningDisabled.com

-----Original Message-----
From: address@hidden
[mailto:address@hidden
On Behalf Of Tamás Nepusz
Sent: Thursday, April 05, 2012 7:35 AM
To: Help for igraph users
Subject: Re: [igraph] Working with large networks and how to sample from a
graph?

> > Thanks
> > 
> > These big networks are hard! My past experience is with networks of 
> > a couple hundred nodes, at most
> > 
Another thing you might want to think about is whether you really need the
betweenness centrality of the nodes or it is enough to use _any_ sensible
centrality measure. For instance, eigenvector centrality and PageRank can be
calculated way faster than betweenness centrality or closeness centrality on
large networks.

T.



_______________________________________________
igraph-help mailing list
address@hidden
https://lists.nongnu.org/mailman/listinfo/igraph-help