View Full Version : how to choose the tools pentaho for my projet?

03-11-2009, 07:10 AM

i has try to put my question here http://forums.pentaho.org/forumdisplay.php?f=87
but nobody can give me an answer. so i try here.

i am not very well in english, so i write my probleme in french, if some people can help me thanks very much. :):)

je suis actuellement travailler dans un sujet de stage:
principale amelioration de la qualité de la notation entre un profil cnadidat et une offre grâce à data mining(weka). à la fin l'outil (pentaho suite) pourra être intégré afin de générer des rapports qui seront utilisées pour détecter les pistes d'amélioration de l'algorithme de scoring.

Pentaho est une plateforme trés riche où l'on peut facilement se perdre. en plus je suis une debutante, donc je me pose la question ici pour discuter avec quelqu'un.

apres je cheche les infos,au debut je pense je dois utilise pentaho data intégration, BIserver, reporting et pentaho data mining(weka). apres je lis plus infos , je me demande est ce que j'ai encore besoin d'autre partie de pentaho ou pas? que je dois choisir pour reporting?(il y a plusieurs ex. report designer, report design wizard etc. si je ne me suis trompe pas.)

merci des connseilles


03-16-2009, 12:29 PM
Helene, I can understand your French, but I'm not very good at writing it and it takes a lot of effort. Hopefully you can understand my English...

If I understand correctly, you want to use a Weka algorithm with some data, and then use the platform to generate reports with the results.

I think you'd be on the right track by using Weka to generate your model, then PDI with the WekaScoring plugin (http://wiki.pentaho.com/download/attachments/1049091/WekaScoring.zip?version=4) to process your data, and finally the BI Server to show reports.

The first part is pretty simple: Weka and PDI. Once you have generated your data you will probably want to use the Pentaho Report Designer (PRD) to create your reports.

Once you have your reports definitions (the xml file generated by PRD), you can use them with the BI platform to display your reports by creating an xaction in Pentaho Design Studio (PDS). You could even automate the process of scoring your data _and_ generating the report, everything in one xaction.

So I would suggest you familiarize yourself with PRD and PDS first. You can find documentation in the wiki:



03-17-2009, 01:53 PM
thanks to answer my question. i thank i can understand your english. :) it's exact what i want in my sujet.
i thank my order is mislead, because i begin with PDI , and now i read mondrian.

so i need :weka, PDI ,Pentaho report designer, pentaho design studio , of course i need BI server. and the first i must to learn reporting designer and design studio.

it's possible that i can contacter with you if i have the question?

thanks again


03-17-2009, 02:17 PM
Well, you would only need Mondrian if you are planning on using a multidimensional database, but that doesn't seem to be the case. I believe you'll do just fine with standard reports, and that's where Pentaho Reporting and the Pentaho Report Designer come in.

Feel free to post your messages in the forums, there's a lot of people here who can help you. You may also send me a private message with a link to your thread in case I miss it and I'll also try to help you out.


03-18-2009, 04:14 AM
ok, thanks