s-news
[Top] [All Lists]

Re: stepAIC in lme models

To: James Rizopoulos <jimsailgr@yahoo.gr>
Subject: Re: stepAIC in lme models
From: Prof Brian Ripley <ripley@stats.ox.ac.uk>
Date: Mon, 31 Dec 2001 10:21:25 +0000 (GMT)
Cc: s-news <s-news@lists.biostat.wustl.edu>
In-reply-to: <20011231081750.5324.qmail@web20107.mail.yahoo.com>
On Mon, 31 Dec 2001, [iso-8859-7] James Rizopoulos wrote:

>
> Dear all,
>
>
>
> My name is James Rizopoulos and I am a new S plus 2000 user for Windows
>
>
>
> I would like to know if it is appropriate to use the stepAIC function of MASS 
> library for an lme model object from the nlme3 library?

No.  There is no extractAIC method for lme models, and by default
lme uses REML, for which there is no AIC.  (Even though one is quoted, it
is not comparable for different fixed effect models.)

One could write a suitable extension of stepAIC, of course.

-- 
Brian D. Ripley,                  ripley@stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272860 (secr)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595


<Prev in Thread] Current Thread [Next in Thread>