You are viewing a plain text version of this content. The canonical link for it is here.

Posted to dev@mahout.apache.org by Raluca Halalai <ra...@yahoo.com> on 2010/05/22 16:22:12 UTC

Community Detection using Mahout

Hello,

First of all, congratulations for a great framework!

My name is Raluca Halalai and I'm a final year undergraduate in CS at the Technical University of Cluj-Napoca, Romania.
I am currently working on my diploma thesis, in community detection.

I used Mahout to implement a community detection algorithm based on the method presented in http://dns2.icar.cnr.it/pizzuti/PPSN08.pdf
It is a genetic algorithm implemented similarly to the Class Discovery evolutionary algorithm .

I tested the community detection method on subsets of data extracted from DBLP http://www.informatik.uni-trier.de/~ley/db/
Due to the size and complexity of DBLP, only a distributed approach is feasible.

If you think this is a useful example for others, I am willing to upload the sources and provide explanations.

Thanks again for a great framework.
Raluca

Re: Community Detection using Mahout

Posted by Robin Anil <ro...@gmail.com>.

Hi Raluca,
                On behalf on the Mahout community, thanks for using Mahout
and glad that you liked it. Code contributions are always welcome. See
https://cwiki.apache.org/MAHOUT/howtocontribute.html on how you can
contribute. Make sure you follow conventions before making the patch. Also
do send a rough diagram/document/email on how it works and how it integrates
with rest of Mahout, with things like Vector, Matrix etc.

Looking forward to your patch
Robin



On Sat, May 22, 2010 at 7:52 PM, Raluca Halalai <ra...@yahoo.com>wrote:

> Hello,
>
> First of all, congratulations for a great framework!
>
> My name is Raluca Halalai and I'm a final year undergraduate in CS at the
> Technical University of Cluj-Napoca, Romania.
> I am currently working on my diploma thesis, in community detection.
>
> I used Mahout to implement a community detection algorithm based on the
> method presented in http://dns2.icar.cnr.it/pizzuti/PPSN08.pdf
> It is a genetic algorithm implemented similarly to the Class Discovery
> evolutionary algorithm .
>
> I tested the community detection method on subsets of data extracted from
> DBLP http://www.informatik.uni-trier.de/~ley/db/
> Due to the size and complexity of DBLP, only a distributed approach is
> feasible.
>
> If you think this is a useful example for others, I am willing to upload
> the sources and provide explanations.
>
> Thanks again for a great framework.
> Raluca
>
>
>