Is this possible?

I am trying to create a mining model so I can predict a
shipment date based on past purchases. Can I have the
shpment date be the predictable column or is that
considered continuous and will not work? Also, I am a
bit confused as to how I would tie together the
purchases, shipments and products tables to reach that
outcome. Any help would be apprciated.

Thanks
Lin

This sounds similar to what ComputerFleet has accomplished in predicting a
return time on rentals. You can see their case study here:
<a style='text-decoration: underline;' href="http://www.microsoft.com/resources/casestudies/CaseStudy.asp?CaseStudyID=14375" target="_blank">http://www.microsoft.com/resources/casestudies/CaseStudy.asp?CaseStudyID=14375</a>

In general I think you want to approach the problem to predict the "days to
ship" rather than the exact "ship date." I.e. "ship date" - "order date".
This number can be discretized and predicted. Your prediction results will
be a range of days representing the time it will take for an order to ship.

Regarding your other questions, I would need more information on your schema
to advise

-Jamie
"Lin Jones" <anonymous.TakeThisOut@discussions.microsoft.com> wrote in message
news:038401c3b38c$96fafb90$a301280a@phx.gbl...
 > I am trying to create a mining model so I can predict a
 > shipment date based on past purchases. Can I have the
 > shpment date be the predictable column or is that
 > considered continuous and will not work? Also, I am a
 > bit confused as to how I would tie together the
 > purchases, shipments and products tables to reach that
 > outcome. Any help would be apprciated.
 >
 > Thanks
 > Lin<!-- ~MESSAGE_AFTER~ -->