Hello,

I'm pretty sure that your edges are weighted because otherwise this
would be an almost complete graph for which there is no sensible
clustering structure anyway. So, I would start by plotting the
distribution of edge weights and throwing away all but the most
heavyweight edges, maybe keeping the top 0.1% only. igraph's
clustering algorithms are designed for sparse graphs, so you would
have to sparsify your graph anyway to do any meaningful analysis on
it.

T.


On Sun, Jul 17, 2016 at 8:08 PM, Jack Zellweger <[email protected]> wrote:
> Hello all,
>
> I hope this email finds you well!
>
> The Problem:
> I am having trouble with run time attempting to cluster a very large
> complete graph of 44 800 nodes and 1 003 497 600 edges.
>
> The Question:
> What is the best igraph function to cluster such a large graph while
> maintaining a good clustering reflecting the true modularity of the graph?
>
> Thank you for your time.
>
> Best,
> Jack Zellweger
> LIGO Research
> Kenyon College
>
> _______________________________________________
> 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

Reply via email to