R help - Nov 2008 - glmer for cauchit link function

Dear all,

A am trying to fit a generalized linear mixed effects model with a binomial
link function, my response data is binary, using the lme4 R package, for the
glmer model but with the cauchit link function (CDF of Cauchy distribution),
under the package this has not yet been coded and was wondering if anyone
knew a way in which I could incorporate this link function into the code.

Thankyou in advance

Liz
-- 
View this message in context:
http://www.nabble.com/glmer-for-cauchit-link-function-tp20612464p20612464.html
Sent from the R help mailing list archive at Nabble.com.

On Thu, Nov 20, 2008 at 5:38 PM, Lizz Metcalfe
<lizz_metcalfe at hotmail.com> wrote:
> I am trying to fit a generalized linear mixed effects model with a binomial
> link function, my response data is binary, using the lme4 R package, for
the
> glmer model but with the cauchit link function (CDF of Cauchy
distribution),
> under the package this has not yet been coded and was wondering if anyone
> knew a way in which I could incorporate this link function into the code.
The current version of lme4 uses C code for the inverse link function
so you would need to modify the function lme4_muEta in the file
lme4/src/lmer.c If you look at the other inverse link functions
defined in that C function you will see that it is not a matter of
simply writing down the inverse link.  You need to be very careful of
the edge conditions and most of the code is devoted to those kinds of
checks.