Akaike Information Criterion is used to aid in model specification -
particularly for determining the number of lags to include. AIC, along with
adjusted R2, Amemiya's prediction criterion and the Schwarz criterion are
different ways to examine the trade off between goodness of fit and parsimony.
One way AIC can be written is: AIC(K)=-2*log-likelihood + 2*K (K is the number
of parameters) where the goal is the find a value of K that minimize AIC.
Scott
----- Original Message -----
From: "Claire Johnstone" <[email protected]>
To: <[email protected]>
Sent: Tuesday, January 14, 2003 10:41 AM
Subject: st: Evaluating Log Likelihood & AIC stat
> Hi,
>
> I have a log likelihood of -2531 for my model. Can I compare this to
> something to evaluate how good this is?
>
> Also, I have an AIC of 0.65 - how do I evaluate this?
>
> Cheers
> Claire
>
> *
> * 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/
