You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@mahout.apache.org by Max <ma...@virgilio.it> on 2009/04/20 11:51:17 UTC

Lucene search result clustering

Hi list,
I would like to do some Lucene Documents clustering.
I have a 
Lucene index and I run my search on the index.
The search result is 
composed of a list of documents.
How can I translate my list of 
document in a format suitable with Mahout format?
I have seen this 
library contains some clustering algorithms, but they don't provide 
(at 
least I haven't found) any translation from a document to a point.
Do I have to implement this by myself, or does it already exist?
Thanks 
in advance.



Re: Lucene search result clustering

Posted by Grant Ingersoll <gs...@apache.org>.
Hi Max,

Great question!  Wish I had a better answer, but unfortunately the  
step from Lucene to Mahout doesn't exist just yet.  I have been slowly  
but surely working on integrating it into Solr (https://issues.apache.org/jira/browse/SOLR-769 
  will eventually have Mahout integration).  I also know we have some  
people on this list working on it: http://www.lucidimagination.com/search/p:mahout?q=Document+clustering 
  but it isn't where it needs to be just yet.

With some luck, there will be a solution soon.  Are you in the  
position to help?

-Grant


On Apr 20, 2009, at 5:51 AM, Max wrote:

> Hi list,
> I would like to do some Lucene Documents clustering.
> I have a
> Lucene index and I run my search on the index.
> The search result is
> composed of a list of documents.
> How can I translate my list of
> document in a format suitable with Mahout format?
> I have seen this
> library contains some clustering algorithms, but they don't provide
> (at
> least I haven't found) any translation from a document to a point.
> Do I have to implement this by myself, or does it already exist?
> Thanks
> in advance.
>
>

--------------------------
Grant Ingersoll
http://www.lucidimagination.com/

Search the Lucene ecosystem (Lucene/Solr/Nutch/Mahout/Tika/Droids)  
using Solr/Lucene:
http://www.lucidimagination.com/search


Re: Lucene search result clustering

Posted by Dawid Weiss <da...@cs.put.poznan.pl>.
If it's just search results clustering and not full document clustering then you 
can take a look at the Carrot2 project. It all depends how you want the 
clustering to be integrated in query processing pipeline, but you can use 
Carrot2 demo applications as a starting point.

http://www.carrot2.org

Dawid

Max wrote:
> Hi list,
> I would like to do some Lucene Documents clustering.
> I have a 
> Lucene index and I run my search on the index.
> The search result is 
> composed of a list of documents.
> How can I translate my list of 
> document in a format suitable with Mahout format?
> I have seen this 
> library contains some clustering algorithms, but they don't provide 
> (at 
> least I haven't found) any translation from a document to a point.
> Do I have to implement this by myself, or does it already exist?
> Thanks 
> in advance.
> 
>