[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: AIC for GAM

From   Yoshiro Nagao <>
Subject   st: AIC for GAM
Date   Tue, 12 Feb 2008 06:09:23 +0900 (JST)

Dear all,

Is it possible to calculate AIC
from an output for "gam"?

"gam" outputs deviance and dispersion.

Although I came across an equation of

AIC = deviance/n + 2*df*deviance/(n-df)/n,

AIC obtained from this calculation was
1000 times larger than the AIC generated
by an equally predictive "glm" model.

Yoshiro Nagao 

Easy + Joy + Powerful = Yahoo! Bookmarks x Toolbar
*   For searches and help try:

© Copyright 1996–2017 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index