specifying random effects covariance structure in nlme

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

specifying random effects covariance structure in nlme

Chris Stride
Hi

I'm trying to fit a mixed effects exponential decay model, in which I
have random effects for the initial value (init), the asymptote (asymp),
and the rate (rate).

The catch is that I'd also like to estimate the correlation between init
and asymp, but not between init and rate, or asymp and rate.

Now usingĀ  random = pdDiag(init + asymp + rate ~ 1) has none of the
random effects correlated

And usingĀ  random = pdSymm(init + asymp + rate ~ 1) has all three of the
random effects correlated

How do I specify just the correlation I want?

cheers

Chris

______________________________________________
[hidden email] mailing list -- To UNSUBSCRIBE and more, see
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.
Reply | Threaded
Open this post in threaded view
|

Re: specifying random effects covariance structure in nlme

Peter Dalgaard-2
I haven't played with this for a decade or so, but I believe you can do something with pdBlocked(). Possibly ask over on R-sig-ME as this quickly gets beyond the R-help level.

(Also, you are aware that it is not about what you want to estimate, but whether you believe the correlations are nonzero?)

-pd

> On 16 Jun 2018, at 13:00 , Chris Stride <[hidden email]> wrote:
>
> Hi
>
> I'm trying to fit a mixed effects exponential decay model, in which I have random effects for the initial value (init), the asymptote (asymp), and the rate (rate).
>
> The catch is that I'd also like to estimate the correlation between init and asymp, but not between init and rate, or asymp and rate.
>
> Now using  random = pdDiag(init + asymp + rate ~ 1) has none of the random effects correlated
>
> And using  random = pdSymm(init + asymp + rate ~ 1) has all three of the random effects correlated
>
> How do I specify just the correlation I want?
>
> cheers
>
> Chris
>
> ______________________________________________
> [hidden email] mailing list -- To UNSUBSCRIBE and more, see
> 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.

--
Peter Dalgaard, Professor,
Center for Statistics, Copenhagen Business School
Solbjerg Plads 3, 2000 Frederiksberg, Denmark
Phone: (+45)38153501
Office: A 4.23
Email: [hidden email]  Priv: [hidden email]

______________________________________________
[hidden email] mailing list -- To UNSUBSCRIBE and more, see
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.