Re: [igraph] Question about the clique percolation method

[Tamas Nepusz]

Dear Frederik,

One thing that you may try is as follows:

1. Upgrade to igraph 0.6 because you will need the maximal.cliques()
function, which is pretty slow in igraph 0.5.*

2. Replace clq <- cliques(graph, min=k, max=k) with:

clq <- maximal.cliques(graph).
clq <- clq[lapply(clq, length) >= k]

3. In the if statement, replace the condition as follows:

length(intersect(clq[[i]], clq[[j]])) >= k-1

This might work; completely untested, but it shows the general idea: instead
of searching for all the k-cliques, we search for the maximal cliques larger
than k. This saves a lot of time if your graph includes lots of large cliques.

-- 
T.

On 04/18/2011 02:20 AM, Frederik Lang wrote:
> Dear all,
> 
> I am new to iGraph (running it in R) and have a question regarding the
> clique percolation function that I found here
> <http://igraph.wikidot.com/community-detection-in-r>.
> 
> I have tried the method on smaller networks (50vertices, 170edges) and it
> has worked fine but when I run it on larger network nothing happens. The
> network I am working with has around 100.000 vertices and 200.000 edges. On
> the website I referred to, it does say the the function is slow but I was
> wondering if being slow means that I just have to try and let it run for
> several hours (overnight) or if it is not possible to use the function on
> such large networks?
> 
> 
> Kind regards,
> 
> 
> Frederik
> 
> 
> 
> _______________________________________________
> igraph-help mailing list
> address@hidden
> http://lists.nongnu.org/mailman/listinfo/igraph-help