Thank you for your comments. They help. One assessment node just compares the two neural nets. The other compares the better neural net with the tree and the logit models and the ensemble. I felt constrained for time and so I didn't extend the project to include much more than minima. I also tried two alternate tree models, one alternate neural net and an alternate logit model, and I discarded them due to flat results. Lift charts attached. db jaison k joseph wrote: Hi Daniel, Why do you have two assessment nodes? You can also try CHAID, CART and C4.5 models. I think, for NN models, input is from Variable selection node. Could you also email me the lift chart? Thanks, Jaison --- Daniel Brockman wrote: > Hi Jaison... > > I solicit your review of my project. With your > permission, I would like the option of including a > copy of your comments with the materials I turn in > for satisfaction of the course requirements. > > Attached is a compressed file containing the > project reports. Extract the contents (download > winzip from http://www.winzip.com/ if needed for the > extraction), then open file em_report.html in your > browser. > > Thank you. > > db > > > DANIEL BROCKMAN 1-707-543-6814 > http://DanielBrockman.com > > > > > > > > > > --------------------------------- > Want to be your own boss? Learn how on Yahoo! Small > Business. DANIEL BROCKMAN 1-707-543-6814 http://DanielBrockman.com --------------------------------- Want to be your own boss? Learn how on Yahoo! Small Business.