The lowest value of AIC indicates always the best model, irrespective of sign. I.e. +5 is better that +10 and -10 is better that -5
S�ren Nielsen
----- Original Message -----
From: "Jonathan A. Schwabish" <[email protected]>
To: <[email protected]>
Sent: Wednesday, January 12, 2005 11:31 PM
Subject: st: AIC question
> I've been playing with the -icomp- user-written
> command, which produces the Akaike Information
> Criterion (AIC). Is it possible for the AIC to be
> negative and if so, is the fit better if the AIC is
> "more negative"? That is, the AIC signals a better fit
> when it is lower, but is the fit better or worse when
> the AIC is -5 versus -10? My understanding would be
> that the regression with an AIC=-10 is a better fit
> but if someone could confirm, I would appreciate it.
>
> Thanks,
> jon
> [email protected]
>
> =====
> Jonathan A. Schwabish, PhD
> Senior Director
> Partnership for New York City
> One Battery Park Plaza
> New York, NY 10004-1479
> www.partnershipfornyc.org
> phone: 212-493-7515
> fax: 212-493-7778
>
>
>
>
> __________________________________
> Do you Yahoo!?
> Read only the mail you want - Yahoo! Mail SpamGuard.
> http://promotions.yahoo.com/new_mail
> *
> * For searches and help try:
> * http://www.stata.com/support/faqs/res/findit.html
> * http://www.stata.com/support/statalist/faq
> * http://www.ats.ucla.edu/stat/stata/
>
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
