Date: Fri, 26 May 2006 07:00:26 -0400
Reply-To: "Miguel ." <statmig@YAHOO.ES>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: "Miguel ." <statmig@YAHOO.ES>
Subject: Re: How Stepwise Regression works in SAS?
I Agree that Stepwise is a technique that should not be used in almost any
kind of Statistical Modelling, and I Think that apart from think a lot about
the problem and about what you trying to predict, descritive statistics is
I Started Working in a Financial Institution recently, and we usually do
Models for Application Scoring. In the first Analysis of the variables, I am
thinking of some Chi-Square Independence tests for the Categorical Variables
vs performance of the Applicant (Good/bad) and some non-parametric bootstrap
model Comparison for the continuous variables vs performance, but I Wonder
if it is really a right approach. Some Comments would be very helpful.