DM: RE: data mining

[Abraham Meidan]

Hi Fatima,
The set of the rules, that are discovered by the data mining 
application, may be viewed as a business model. You can check the 
accuracy of this business model by issuing predictions on your data. 
Divide your data into two parts, one part will be used as a training 
file, where the data mining application will reveal the rules, and 
the other part will be the test file, where the data mining 
application will issue the predictions. You will then check how 
accurate the predictions are (in comparison with the actual data). As 
long as they are accurate, you have a reliable business model.
According to my experience, you get the best business model through 
association rules (the set of ALL if-then rules).
Best regards,
Abraham.

-----Original Message-----
From:   Fatima Goncalves [SMTP:mflfag@mat.uc.pt]
Sent:   ב פברואר 16 1998 12:22
To:     datamine-l@nautilus-sys.com
Subject:        DM: data mining

I am interested in something with data mining and business model or 
DSS.
I do not know if there is any papers in these subjects.
Everything I read says:
"You do data mining and you have the rules; so, you "take" the rules 
and you modify your model"
I do not want this. I want to confront my business knowledge with 
data mining and to obtain knew knoeledge to change (or not...) the 
business model.
Can you help me?
Fatima