Fitting multi-response mixed effects models with lmer

4 messages · jonnations, Ian Dworkin, Ben Bolker

Original

1

4

jonnations

Sun, Jul 14, 2019 7:32 AM #

Hi Alejandro,

This is easy to do in brms, if you?re willing to explore Bayesian options.
There is a nice vignette (brms multivariate) that covers this exact thing.

Jon

On Sun, Jul 14, 2019 at 3:01 AM <r-sig-mixed-models-request at r-project.org>
wrote:

        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. Fitting multi-response mixed effects models with lmer
      (Alejandro Catalina)

----------------------------------------------------------------------

Message: 1
Date: Sat, 13 Jul 2019 18:02:20 +0300
From: Alejandro Catalina <alecatfel at gmail.com>
To: r-sig-mixed-models at r-project.org
Subject: [R-sig-ME] Fitting multi-response mixed effects models with
        lmer
Message-ID: <e44b3683-1a88-45f8-8dc8-1c07595c0dd7 at Spark>
Content-Type: text/plain; charset="utf-8"

Dear all,

I found myself trying to fit a multi-response model with lmer the other
day and today I learned that it is indeed not implemented. Is there anyone
looking on that direction or does anyone have any pointers or suggestions?
I guess I can iteratively fit one model for each response but I?m guessing
that would be much slower. Furthermore, I would need to later combine all
the models into a single object for my specific requirements. This is the
issue I opened on lme4?s GitHub:

        Hi,
I am trying to solve the following formula with lmer:
cbind(y.1, y.2, y.3) ~ u + (u | floor_id) + (u | county_id)
which works fine for standard lm models without the group terms, but it
fails when I have the mixed effects terms with the following error:
Error in initializePtr() : updateMu: Size mismatch
If this is not the right place to post this issue please tell me, I
appreciate any pointers forward.

Thank you all,

Best,
Alejandro

        [[alternative HTML version deleted]]




------------------------------

Subject: Digest Footer

_______________________________________________
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 151, Issue 10
***************************************************

Jonathan A. Nations
PhD Candidate
Esselstyn Lab <https://esselstyn.github.io/>
Museum of Natural Sciences <https://www.lsu.edu/mns/>
Louisiana State University

	[[alternative HTML version deleted]]

Ben Bolker

Sun, Jul 14, 2019 1:07 PM #

It isn't terribly hard to roll your own: this is untested but should
get you started.

   respvars <- c("y.1","y.2","y.3")
   fits <- vector("list", 3)
   names(fits) <- respvars
   fits[[1]] <- lmer(y.1 ~ u + (u | floor_id) + (u | county_id),
data=your_data))
   for (i in 2:3) {
           fits[[i]] <- refit(fits[[1]], your_data[[respvars[i]]]
       }
  }

On Sun, Jul 14, 2019 at 10:33 AM jonnations <jonnations at gmail.com> wrote:

        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. Fitting multi-response mixed effects models with lmer
      (Alejandro Catalina)

----------------------------------------------------------------------

Message: 1
Date: Sat, 13 Jul 2019 18:02:20 +0300
From: Alejandro Catalina <alecatfel at gmail.com>
To: r-sig-mixed-models at r-project.org
Subject: [R-sig-ME] Fitting multi-response mixed effects models with
        lmer
Message-ID: <e44b3683-1a88-45f8-8dc8-1c07595c0dd7 at Spark>
Content-Type: text/plain; charset="utf-8"

Dear all,

I found myself trying to fit a multi-response model with lmer the other
day and today I learned that it is indeed not implemented. Is there anyone
looking on that direction or does anyone have any pointers or suggestions?
I guess I can iteratively fit one model for each response but I?m guessing
that would be much slower. Furthermore, I would need to later combine all
the models into a single object for my specific requirements. This is the
issue I opened on lme4?s GitHub:

        Hi,
I am trying to solve the following formula with lmer:
cbind(y.1, y.2, y.3) ~ u + (u | floor_id) + (u | county_id)
which works fine for standard lm models without the group terms, but it
fails when I have the mixed effects terms with the following error:
Error in initializePtr() : updateMu: Size mismatch
If this is not the right place to post this issue please tell me, I
appreciate any pointers forward.

Thank you all,

Best,
Alejandro

        [[alternative HTML version deleted]]




------------------------------

Subject: Digest Footer

_______________________________________________
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 151, Issue 10
***************************************************

_______________________________________________
R-sig-mixed-models at r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models

Ian Dworkin

Sun, Jul 14, 2019 1:36 PM #

Alejandro,

 Ben B. and I taught some examples of "tricking" lmer for multivariate
response models, see here
https://mac-theobio.github.io/QMEE/MultivariateMixed.html

Cheers
Ian

On Sun, 14 Jul 2019 at 10:33, jonnations <jonnations at gmail.com> wrote:

        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. Fitting multi-response mixed effects models with lmer
      (Alejandro Catalina)

----------------------------------------------------------------------

Message: 1
Date: Sat, 13 Jul 2019 18:02:20 +0300
From: Alejandro Catalina <alecatfel at gmail.com>
To: r-sig-mixed-models at r-project.org
Subject: [R-sig-ME] Fitting multi-response mixed effects models with
        lmer
Message-ID: <e44b3683-1a88-45f8-8dc8-1c07595c0dd7 at Spark>
Content-Type: text/plain; charset="utf-8"

Dear all,

I found myself trying to fit a multi-response model with lmer the other
day and today I learned that it is indeed not implemented. Is there

_______________________________________________
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 151, Issue 10
***************************************************

_______________________________________________
R-sig-mixed-models at r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models

Ian Dworkin
Department of Biology
McMaster University
Office phone 905 525 9140 ext. 21775
Lab phone 905 525 9140 ext. 20076
dworkin at mcmaster.ca

	[[alternative HTML version deleted]]

Ben Bolker

Sun, Jul 14, 2019 2:03 PM #

I thought Alejandro was interested in fitting these responses as
*independent* outcomes (since he says below "I can iteratively fit one
model for each response but I?m guessing that would be much slower"; I
think the loop using refit() every time after the first would be
reasonably fast - I certainly don't see a super-easy way to do it faster
...)

On 2019-07-14 4:36 p.m., Ian Dworkin wrote:

        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. Fitting multi-response mixed effects models with lmer
      (Alejandro Catalina)

----------------------------------------------------------------------

Message: 1
Date: Sat, 13 Jul 2019 18:02:20 +0300
From: Alejandro Catalina <alecatfel at gmail.com>
To: r-sig-mixed-models at r-project.org
Subject: [R-sig-ME] Fitting multi-response mixed effects models with
        lmer
Message-ID: <e44b3683-1a88-45f8-8dc8-1c07595c0dd7 at Spark>
Content-Type: text/plain; charset="utf-8"

Dear all,

I found myself trying to fit a multi-response model with lmer the other
day and today I learned that it is indeed not implemented. Is there

_______________________________________________
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 151, Issue 10
***************************************************

_______________________________________________
R-sig-mixed-models at r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models