|Date: ||Sat, 14 Jun 2008 11:03:39 -0400|
|Reply-To: ||Peter Flom <email@example.com>|
|Sender: ||"SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>|
|From: ||Peter Flom <peterflomconsulting@MINDSPRING.COM>|
|Subject: ||Re: Outlier effect|
|Content-Type: ||text/plain; charset=UTF-8|
Abhay Kaushik <abhaykaushik@HOTMAIL.COM> wrote
>I am running a multivariate regression where two of the variables have
>extreme outliers. What options I can use to eliminate or minimize their
>effect on my overall results.
Well, first I'd want to know *why* these variables have outliers.
1) Deleting the outliers
2) Transforming the variables
3) Using robust forms of regression
4) Using a regression tree
and probably other things as well.... without some context, it's hard to give a recommendation.
See my post on sas-community, How to ask a statistics question
Peter L. Flom, PhD
www DOT peterflom DOT com