[edm-discuss] Re: Clustering with more DB tables

  • From: Andreza Vieira <andreza.ufcg@xxxxxxxxx>
  • To: edm-discuss@xxxxxxxxxxxxx
  • Date: Wed, 24 Jun 2009 12:57:12 -0300


This single Database Table is just a text file according to ARFF
There's several tools that support this database format, as WEKA (this tool
also supports other database formats).

If you have a database with 65 tables, you should agglutinate all these
tables in only one, a big table (ARFF). For this, you can make *joins *between
these tables.

The database format that you will need depends on the database tool that you
will use. What's the tool that you will use?


2009/6/24 qazmlp q <qazmlp1209@xxxxxxxxxxxxxx>

> Hi,
> I'm relatively new to the Data mining.
> Whatever the online tutorials and the books that I checked, explain the
> Data mining concepts using a single Database Table.
> But practically, there will be around 50-100 tables for typical projects.
> I'm not clear how do we have to use these tools in such cases?
> For example, I will have to perform clustering for all the data available
> in the Database, and the Database contain around 65 tables. I'm sure that
> this is a very typical case for all. How do I have to proceed in such cases?
> Please elaborate if there are any preprocessing steps that are necessary.
> Regards,
> Micheal
> <http://sigads.rediff.com/RealMedia/ads/click_nx.ads/www.rediffmail.com/signatureline.htm@Middle?>

Andreza Vieira
Computer Science MSc Student
Formal Methods Group - http://www.gmf.ufcg.edu.br
Systems and Computing Department - http://www.dsc.ufcg.edu.br
Center of Electrical Engineering and Informatics - http://ceei.ufcg.edu.br
Federal University of Campina Grande - http://ufcg.edu.br

Other related posts: