[R] lme correlation structures

[Gareth Hughes]

Thanks for your help. I can weight the variances in the 2 groups using
"weights", as in

lme(Y~1+time+sex+age, random=~1|indv,
correlation=corAR1(form=~time|indv),
weights=varIdent(form=~1|sex),method="ML")

but what I would like is to have a different "phi" estimate for each
gender, not just different variances. Will track down a copy of the
Pinheiro and Bates book...



On 6/27/07, Bert Gunter <gunter.berton at gene.com> wrote:
> Please read ?lme carefully -- the info you seek is there. In particular, the
> weights argument for changing variance weighting by covariates and the
> correlation argument for specifying correlation structures.
>
> Pinheiro and Bates's MIXED EFFECT MODELS IN S... is the canonical reference
> (which you should get if you want to use R as you said) that exposits the
> ideas at greater length.
>
>
> Bert Gunter
> Genentech Nonclinical Statistics
>
> -----Original Message-----
> From: r-help-bounces at stat.math.ethz.ch
> [mailto:r-help-bounces at stat.math.ethz.ch] On Behalf Of Gareth Hughes
> Sent: Wednesday, June 27, 2007 7:50 AM
> To: r-help at stat.math.ethz.ch
> Subject: [R] lme correlation structures
>
> Hi all,
>
> I've been using SAS proc mixed to fit linear mixed models and would
> like to be able to fit the same models in R. Two things in particular:
>
> 1) I have longitudinal data and wish to allow for different repeated
> measures covariance parameter estimates for different groups (men and
> women), each covariance matrix having the same structure. In proc
> mixed this would be done by specifying group= in the REPEATED
> statement. Is this simple to do in R? (I've tried form=~time|indv/sex
> for example but this doesn't seem to do the job).
>
> 2) I've read that other correlation structures can be specified. Does
> anyone have any examples of how toeplitz or (first-order)
> ante-dependence structures can be specified?
>
> Many thanks,
>
> Gareth
>
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>
>