Re: [igraph] 'decompose.graph' versus 'clusters'

[Gabor Csardi]

Oh, that is very good, I believe that the cutoff=5 estimation must 
be already very close to the real value in most real graphs
(real graph=diameter is small, so all shortest paths are short).

Did you try to compare cutoff=4 and cutoff=5 results? If they are 
about the same that is a good sign.

Gabor

On Fri, Jul 25, 2008 at 08:23:04AM -0700, David Hunkins wrote:
> And by the way, betweenness.estimate(G,cutoff=5) worked on my weakly  
> connected graph G with 2M vertices and 2M edges. It only took two hours 
> on the 'fast' EC2 machine. Thanks again,
>
> Dave
>
> David Hunkins
> address@hidden
> im: davehunkins
> 415 336-8965
>
[...]

-- 
Csardi Gabor <address@hidden>    UNIL DGM