Hi all,

There is already a mailinglist for data mining and data mining techniques (like AT's techniques). you can subscribe by filling out this form: http://www.nautilus-sys.com/form1.html

The list is not 'SPSS specific' and not nearly as busy as the SPSS list but you can get some good answers to your questions, especially the questions with a more technical background.

Oh, by the way, we didn't promise AT v2 will be 50x faster, it is 'only' at least 5x faster ;-)

Martijn Wiertz SPSS Benelux

-----Original Message----- From: Vlad Balaban [SMTP:VladB@POLK.CA] Sent: Monday, April 12, 1999 4:25 PM To: SPSSX-L Subject: Re: Answer Tree

I have just run a large job with Answer Tree 1.0. I loaded in 220,000 cases and asked it to build a tree on 65 categories of a dependent variable. I used exhaustive CHAID which seemed giving best results and kept the default 100/50 node sizes.

I run the job overnight (understand that Answer Tree 2.0 is 50 times faster). However, the huge graphic image I created was too much for my computer. I obtanied a tree with about 2,800 nodes that my 64M RAM had hard time displaying. I wished there was a way to open it without being thrown into the graphic screen. If I could not see the tree, I could at least read the gains and risk statistics. But there did not seem to be any way around that huge graphic image when opening it.

I may still be doing additionnal work on this project and I will enjoy reading your experiences.

Vlad

*********************************************************** Dr. Vladislav Balaban Research Consultant Compusearch Custom Research Polk Marketing Canada

> -----Original Message----- > From: Dean [SMTP:ombudsman@POBOX.COM] > Sent: Sunday, April 11, 1999 8:23 AM > To: SPSSX-L@UGA.CC.UGA.EDU > Subject: Answer Tree > > Are there any specific groups for Answer Tree? > > Does anyone out there use Answer Tree - Have any feedback about what > they think > about the product? > > TIA > Dean