| Date: | Mon, 15 Apr 1996 12:12:10 -0500 |
|---|
| Reply-To: | "Scot A. Richardson" <cma-mkt@INTERACCESS.COM> |
|---|
| Sender: | "SAS(r) Discussion" <SAS-L@UGA.CC.UGA.EDU> |
|---|
| From: | "Scot A. Richardson" <cma-mkt@INTERACCESS.COM> |
|---|
| Organization: | Conway/Milliken & Associates |
|---|
| Subject: | Re: Variance Inflation Factors |
|---|
ko02@Lehigh.EDU wrote:
>
> Hello:
>
> I would like to know if PROC GLM or PROC REG
> have options for computing VIF.
>
> Any help is greatly appreciated.
>
> Sincerely,
>
> Kouros Owzar.
You could just try:
proc reg data=yourdata.set;
model y=x1-x3 /tol;
run;
Where tol is the reciprocal of vif or... just go at it directly
(grin--sorry about that tolerance thing...just couldn't help myself)
proc reg data=yourdata.set;
model y=x1-x3 /vif;
run;
Take a peek at SAS/Stat Vol II page 1368-69! Good luck
Scot A. Richardson
Project Mgr Statistical Services
UNIX System Admin
Conway/Milliken & Associates
875 N. Michigan Ave.
John Hancock Center Suite 2511
Chicago, IL 60611
| 
