igraph-help
[Top][All Lists]

## Re: [igraph] Choosing between different methods of detecting communities

 From: 凌琛 Subject: Re: [igraph] Choosing between different methods of detecting communities Date: Thu, 4 Oct 2012 10:24:44 +0800

Hi Gabor,

Sorry, I went to sleep last night.

I attached the codes and results here.

> G = read.graph("D:/text mining project/term lists/40 core term list/GraphML/50w_lift.graphml",  "graphml")
> ec <- edge.betweenness.community(G, E(G)\$weight, FALSE)
> ec
Graph community structure calculated with the edge betweenness algorithm
Number of communities (best split): 15
Modularity (best split): 0.02788969
Membership vector:
[1]  1  2  3  4  5  2  2  2  6  7  2  8  2  2  2  2  2  2  2  9  2  2  2  2  2  2  2 10 11  2 12  2 13  2  2 14  2  2 15  2
> ec <- edge.betweenness.community(G, NULL, FALSE)
> ec
Graph community structure calculated with the edge betweenness algorithm
Number of communities (best split): 1
Modularity (best split): 0
Membership vector:
[1] 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

Did I do something wrong here? The source data is in the attachment. Thanks.

Best Regards,
chen

On Wed, Oct 3, 2012 at 11:44 PM, Gábor Csárdi wrote:
If you get just one community, then you are surely doing something
wrong (or there is a bug in igraph).

edge.betweenness.community is working fine for our test cases, so
you'll need to show us a reproducible example that does not work the
way you expect it to.

Gabor

On Wed, Oct 3, 2012 at 11:34 AM, 凌琛 <address@hidden> wrote:
> yes, maximun the modularity.
>
> Actually when the result is only one community, the modularity is 0.
>
> SNAP is a small library developed by Stanford. You can take a look when you
> have time. Igraph is much more complete, thanks for the development and
> sharing.
>
> Regards,
> chen
>
> On Oct 3, 2012 11:14 PM, "Tamás Nepusz" <address@hidden> wrote:
>>
>> > I know that the algorithm is not deterministic. In my case, the graph
>> > only have tens of nodes, while the results are very different.
>> > In igraph, the unweighted version results in only one community; in
>> > SNAP, there are quite a few communities.
>>
>> How does SNAP select the number of communities for the edge betweenness
>> method? This is not specified in the original algorithm; igraph just cuts
>> the community dendrogram at the point where the modularity is maximal. Is it
>> the same for SNAP?
>>
>> Cheers,
>> T.
>>
>>
>>
>> _______________________________________________
>> igraph-help mailing list
>> https://lists.nongnu.org/mailman/listinfo/igraph-help
>
>
> _______________________________________________
> igraph-help mailing list
> https://lists.nongnu.org/mailman/listinfo/igraph-help
>

--
Gabor Csardi <address@hidden>     MTA KFKI RMKI

_______________________________________________
igraph-help mailing list