[R-sig-ME] R questions
John Maindonald
john.maindonald at anu.edu.au
Fri Aug 3 14:11:45 CEST 2007
Whether or not the BLUPs are very similar to the means depends
on the relative magnitudes of the relevant components of variance.
Try the following code (from p.309 of the 2nd edn of my book with
John Braun):
ant111b.lmer <- lmer(harvwt ~ 1 + (1 | site), data=ant111b)
options(digits=4)
ant111b.lmer
## Fitted values and residuals in \texttt{lmer()}
s2W <- 0.578; s2L <- 2.37; n <- 4
sitemeans <- with(ant111b, sapply(split(harvwt, site), mean))
grandmean <- mean(sitemeans)
shrinkage <- (n*s2L)/(n*s2L+s2W)
grandmean + shrinkage*(sitemeans - grandmean)
##
## More directly, use fitted() with the lmer object
unique(fitted(ant111b.lmer))
##
## Compare with site means
sitemeans
John Maindonald email: john.maindonald at anu.edu.au
phone : +61 2 (6125)3473 fax : +61 2(6125)5549
Centre for Mathematics & Its Applications, Room 1194,
John Dedman Mathematical Sciences Building (Building 27)
Australian National University, Canberra ACT 0200.
On 3 Aug 2007, at 5:46 PM, Iasonas Lamprianou wrote:
> Dear friends,
>
> in this web page http://cran.r-project.org/src/contrib/Views/
> Psychometrics.html, I read: A multilevel Rasch model can be
> estimated using the package lme4 with functions for mixed-effects
> models with crossed or partially crossed random effects. How can
> this be done? Is there an example to read?
>
> Also, why are the random effects (BLUPS usinf the ranef fucnction)
> and the fixed effects estimates the same most pf the time?
>
> Thanks
>
> jason
>
> Dr. Iasonas Lamprianou
> Department of Education
> The University of Manchester
> Oxford Road, Manchester M13 9PL, UK
> Tel. 0044 161 275 3485
> iasonas.lamprianou at manchester.ac.uk
>
>
> ----- Original Message ----
> From: "r-sig-mixed-models-request at r-project.org" <r-sig-mixed-
> models-request at r-project.org>
> To: r-sig-mixed-models at r-project.org
> Sent: Thursday, 2 August, 2007 1:00:22 PM
> Subject: R-sig-mixed-models Digest, Vol 8, Issue 1
>
>
> Send R-sig-mixed-models mailing list submissions to
> r-sig-mixed-models at r-project.org
>
> To subscribe or unsubscribe via the World Wide Web, visit
> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
> or, via email, send a message with subject or body 'help' to
> r-sig-mixed-models-request at r-project.org
>
> You can reach the person managing the list at
> r-sig-mixed-models-owner at r-project.org
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of R-sig-mixed-models digest..."
>
>
> Today's Topics:
>
> 1. simulate() fails for poisson lmer fits (William Valdar)
> 2. Re: simulate() fails for poisson lmer fits (William Valdar)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Wed, 1 Aug 2007 18:49:58 +0100 (BST)
> From: William Valdar <valdar at well.ox.ac.uk>
> Subject: [R-sig-ME] simulate() fails for poisson lmer fits
> To: R-SIG-Mixed-Models <r-sig-mixed-models at r-project.org>
> Message-ID: <Pine.LNX.4.64.0708011822100.8824 at zeon.well.ox.ac.uk>
> Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
>
> Hello,
>
> I wish to simulate() from a fitted Poisson GLMM. Both lmer() and
> lmer2()
> from lme4 (version info at the bottom) fail, apparently due to bugs.
> Here's a test case:
>
> counts <- data.frame(
> count = c(8, 3, 0, 3, 0, 9, 0, 11, 4, 7, 0, 0, 0, 4, 3,
> 6, 3,
> 15, 11, 9),
> batch = c(3.1, 3.1, 3.1, 3.3, 3.3, 3.3, 3.2, 3.12, 3.8, 3.11,
> 3.4, 3.4, 3.4, 3.4, 3.5, 3.5, 3.5, 3.5, 3.6, 3.6),
> weight = c(324.4, 372.5, 352.7, 379.6, 388.1, 431, 448.4,
> 377.3,
> 376.5, 358.4, 356, 351.4, 350.8, 332.1, 334.5,
> 392, 370.5,
> 409.7, 375, 318.5))
>
> fit <- lmer(count ~ weight + (1|batch), family=poisson, data=counts)
> # Error: inherits(object, "lmer") is not TRUE
>
> fit <- lmer2(count ~ weight + (1|batch), family=poisson, data=counts)
> simulate(fit)
> # CHOLMOD error: X and/or Y have wrong dimensions
> # Error in crossprod(object at ZXyt, c(unlist(lapply(seq_along(re),
> # function(k) (t(cholL[[k]]) %*% :
> # Cholmod error 'X and/or Y have wrong dimensions' at
> # file:../MatrixOps/cholmod_sdmult.c, line 90
>
> Is there a quick fix for either of these two? Otherwise, is there an
> alternative (I've checked objects produced by nlme, glmmPQL, GLMMGibbs
> with no luck)? I am using lme4 0.99875-6 version R 2.5.1 on Windows
> XP.
>
> Many thanks,
>
> Will
>
> =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
> Dr William Valdar ++44 (0)1865 287 589
> Wellcome Trust Centre valdar at well.ox.ac.uk
> for Human Genetics, Oxford www.well.ox.ac.uk/~valdar
>
>
>
> ------------------------------
>
> Message: 2
> Date: Thu, 2 Aug 2007 00:23:07 +0100 (BST)
> From: William Valdar <valdar at well.ox.ac.uk>
> Subject: Re: [R-sig-ME] simulate() fails for poisson lmer fits
> To: R-SIG-Mixed-Models <r-sig-mixed-models at r-project.org>
> Message-ID: <Pine.LNX.4.64.0708020021210.7087 at zeon.well.ox.ac.uk>
> Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
>
>
> I missed out a line in the code below. The line that throws the
> error is
> not the fitting but the subsequent simulation, ie:
>
> fit <- lmer(count ~ weight + (1|batch), family=poisson, data=counts)
> simulate(fit)
> # Error: inherits(object, "lmer") is not TRUE
>
>
> On Wed, 1 Aug 2007, William Valdar wrote:
>> Hello,
>>
>> I wish to simulate() from a fitted Poisson GLMM. Both lmer() and
>> lmer2() from
>> lme4 (version info at the bottom) fail, apparently due to bugs.
>> Here's a test
>> case:
>>
>> counts <- data.frame(
>> count = c(8, 3, 0, 3, 0, 9, 0, 11, 4, 7, 0, 0, 0, 4, 3, 6, 3,
>> 15, 11, 9),
>> batch = c(3.1, 3.1, 3.1, 3.3, 3.3, 3.3, 3.2, 3.12, 3.8, 3.11,
>> 3.4, 3.4, 3.4, 3.4, 3.5, 3.5, 3.5, 3.5, 3.6, 3.6),
>> weight = c(324.4, 372.5, 352.7, 379.6, 388.1, 431, 448.4,
>> 377.3,
>> 376.5, 358.4, 356, 351.4, 350.8, 332.1, 334.5, 392,
>> 370.5,
>> 409.7, 375, 318.5))
>>
>> fit <- lmer(count ~ weight + (1|batch), family=poisson, data=counts)
>> # Error: inherits(object, "lmer") is not TRUE
>>
>> fit <- lmer2(count ~ weight + (1|batch), family=poisson,
>> data=counts)
>> simulate(fit)
>> # CHOLMOD error: X and/or Y have wrong dimensions
>> # Error in crossprod(object at ZXyt, c(unlist(lapply(seq_along(re), #
>> function(k) (t(cholL[[k]]) %*% :
>> # Cholmod error 'X and/or Y have wrong dimensions' at #
>> file:../MatrixOps/cholmod_sdmult.c, line 90
>>
>> Is there a quick fix for either of these two? Otherwise, is there an
>> alternative (I've checked objects produced by nlme, glmmPQL,
>> GLMMGibbs with
>> no luck)? I am using lme4 0.99875-6 version R 2.5.1 on Windows XP.
>>
>> Many thanks,
>>
>> Will
>>
>> =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
>> Dr William Valdar ++44 (0)1865 287 589
>> Wellcome Trust Centre valdar at well.ox.ac.uk
>> for Human Genetics, Oxford www.well.ox.ac.uk/~valdar
>>
>>
>
> =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
> Dr William Valdar ++44 (0)1865 287 589
> Wellcome Trust Centre valdar at well.ox.ac.uk
> for Human Genetics, Oxford www.well.ox.ac.uk/~valdar
>
>
>
> ------------------------------
>
> _______________________________________________
> R-sig-mixed-models mailing list
> R-sig-mixed-models at r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
>
>
> End of R-sig-mixed-models Digest, Vol 8, Issue 1
> ************************************************
>
>
> ___________________________________________________________
>
> _______________________________________________
> R-sig-mixed-models at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
More information about the R-sig-mixed-models
mailing list