[prev in list] [next in list] [prev in thread] [next in thread] 

List:       r-sig-mixed-models
Subject:    [R-sig-ME] lme - incorporating measurement error with estimated V-C
From:       Todd Ogden <to166 () columbia ! edu>
Date:       2010-02-19 16:23:10
Message-ID: B14203EC-5ACE-4873-907E-EE447D039ED6 () columbia ! edu
[Download RAW message or body]

I have data (each Y_i is a vector) in the form of

Y_i = X_i \beta_i   + Z_i b_i     + epsilon_i

Were it not for the measurement error (the epsilon_i) it's a very
simple model --- nice and balanced, compound symmetry, and I'd just
use lme(y ~ x1 + x2, random=~1|subj, ...) but the measurement error is
throwing me off.

Because the Y_i are actually derived from other data, I am able to do
some bootstrapping and get an estimate of the V-C matrix of epsilon_i.

But I haven't been able to figure out how to weight the observations
properly in an lme() call.

Some searching of the archives led me to a 2004 posting (courtesy of
Dave Atkins) of two functions written by Jose: varRan and varWithin.
This gives me some hope (a good deal of hope, actually), but I can't
understand the arguments or how to use these functions.

Here's the posting:
http://tolstoy.newcastle.edu.au/R/help/04/04/0245.html

Any hints would be greatly appreciated.

Thanks,

Todd


	[[alternative HTML version deleted]]

_______________________________________________
R-sig-mixed-models@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
[prev in list] [next in list] [prev in thread] [next in thread]