s-news
[Top] [All Lists]

NLME plotting and Confidence Intervals

To: Splus Listserv <s-news@lists.biostat.wustl.edu>
Subject: NLME plotting and Confidence Intervals
From: Greg Tarpinian <sasprog474@yahoo.com>
Date: Thu, 18 Nov 2004 18:11:14 -0800 (PST)
Comment: DomainKeys? See http://antispam.yahoo.com/domainkeys
Domainkey-signature: a=rsa-sha1; q=dns; c=nofws; s=s1024; d=yahoo.com; b=lFerbjyhyFb2GbsWMf50MZsllz/1fDywmKZHtTvyIdJwu23VBjtWA7v/xqSzr70+o1SLSA13X7hdJ9bgY2hKVGYVoS04jVoZ3wp00XBS19K3Pgn1hZC7MbntPp6JU0O4j7MFBz0xAdhA2nk0n570T1uQWWJ+/Bt5yoZppceHGiw= ;
All,

(I have also posted this to the R-help listserv as I 
am using both R and Splus.)  I am a relative newcomer
to Splus and have been learning about the NLME 
package and mixed modeling.  I have been able to 
successfully fit some simple linear and 4PL logistic
models using lme( ) and nlme( ).  My questions are

(1) How can I import a SAS dataset with DATE9.    
    formatted
    date variable and get this information into a
    convenient format for use with the trellis graphs?
    In particular, I would to produce panel plots of
    the sort that Pinheiro and Bates discuss in their
    very helpful book.

(2) How can I plot the fitted response function along
    with 95% confidence intervals around the estimated
    mean response?  I suspect this is harder to do
with
    nonlinear mixed models, and I would really 
    appreciate some guidance here.

(3) How can I find and plot prediction intervals for
    individual subjects?

Sorry for what must be trivial questions!  I really
appreciate any help.

Thanks,
     Greg


                
__________________________________ 
Do you Yahoo!? 
The all-new My Yahoo! - Get yours free! 
http://my.yahoo.com 
 


<Prev in Thread] Current Thread [Next in Thread>
  • NLME plotting and Confidence Intervals, Greg Tarpinian <=