View Full Version : Where is Pentaho Data Mining functionality?

01-07-2008, 10:46 AM
Hi, wanted to start testing Pentaho Data Mining. I thought there would be samples available inside last version of Pentaho-demo, but I cannot see them. And I also thought Pentaho Data Mining would have all Weka functionality integrated first inside Pentaho SDK, and then, later on, would be made available through Pentaho Design Studio (just like reports, execution of Kettle jobs...).

I've downloaded Pentaho 1.6 GA 863 version, both the demo and the SDK, but cannot see anything related to Pentaho Data Mining. Where I can fin Weka related classes inside Pentaho SDK?

Or, means this I should use Weka on my own, built a completely appart Java application and then call this application from Pentaho?

I would really appreciate any help on getting started with this new Pentaho module.

Thanks very much in advance.

01-07-2008, 04:46 PM
Hi Miguel,

Work on integration of Weka into the Pentaho platform is in its nascent stages. At present, there are some plugins for Kettle that allow pre-build Weka classifiers and clusterers to be used to score data in a transform. More integration should occur this year.

Best regards,

01-08-2008, 07:20 AM
Thanks Marc, it's crystal clear now that if we want to use Weka from Pentaho, we must add the corresponding .jars to Pentaho SDK and develop a custom component (.java) that plays with Weka functionality.