Re: [igraph] graph partitioning in K groups

[Claudio Martella]

Hi Gabor,

thanks for answering. I'm not sure i understand exactly what you mean.
I use python, but finding k-means for python would not be a problem,
the problem would be using k-means itself for graph-partitioning? what
would the feature vector be? the neighborhood?

i'm actually looking for something more suited for graphs, as you
suggest to look at community detection. By topology I in reality mean
connectivity-based graph partitioning, something that puts vertices in
cluster according to the fact they are similarly connected (in this
sense it's not SO different from the previous idea of k-means on
neighborhoods) but in a "sophisticated" way.

I'll look at the two algorithms that you suggest.

thanks!

On Mon, Dec 19, 2011 at 4:38 PM, Gábor Csárdi <address@hidden> wrote:
> Hi,
>
> if you use R, you can use one of the built-in clustering functions,
> e.g. kmeans().
>
> But actually you can partition a graph into a given number of groups
> with many of the community finding algorithms, because they return a
> complete dendrogram. E.g. the fast greedy and the edge betweenness
> based algorithms are like that. This might not be what you want
> though, depending on what exactly you mean by 'topologically'.
>
> Gabor
>
> On Mon, Dec 19, 2011 at 6:16 AM, Claudio Martella
> <address@hidden> wrote:
>> Hello,
>>
>> I need to partition a weighted undirected graph topologically in a set
>> of K groups, something like k-means. all i can find is
>> community-detection without the possibility to specify K.
>> Can anybody give me a reference for a way of doing this in igraph or
>> in case to an algorithm that does so I can implement myself?
>>
>> thanks!
>>
>> --
>>    Claudio Martella
>>   address@hidden
>>
>> _______________________________________________
>> igraph-help mailing list
>> address@hidden
>> https://lists.nongnu.org/mailman/listinfo/igraph-help
>
>
>
> --
> Gabor Csardi <address@hidden>     MTA KFKI RMKI
>
> _______________________________________________
> igraph-help mailing list
> address@hidden
> https://lists.nongnu.org/mailman/listinfo/igraph-help



-- 
   Claudio Martella
  address@hidden