[R] logit link + alternatives

[Martyn Plummer]

I am not sure that fully answers Jeff's question.  If the available link
functions (even with the quasi family) are not sufficient for your
needs, then you need to make your own constructor for a "family" object
- which provides the necessary information to the glm engine - and use
this as the family argument in your call to glm.  Specifically, you need
to make a copy of the binomial() function and replace the call to
make.link().

Martyn

On Mon, 2005-02-07 at 11:27 +0100, Dimitris Rizopoulos wrote:
> see at ?glm and ?family
> 
> and use, e.g., `family=binomial(link="probit")'.
> 
> Best,
> Dimitris

> ----- Original Message ----- 
> From: <jeff.pr2 at added-insight.net>
> To: <r-help at stat.math.ethz.ch>
> Sent: Monday, February 07, 2005 10:44 AM
> Subject: [R] logit link + alternatives
> 
> 
> >
> > Help needed with lm function:
> >
> > Dear R's,
> > Could anyone tell me how to replace the link function (probit logit,
> > loglog etc.) in lm
> > with an abitrary user-defined function? The task is to perform ML
> > Estimation of betas
> > for a dichotome target variable.
> >
> > Maybe there is already a package for this (I did not find one).
> > Any hints or a code excerpt would be welcome!
> > Thank you -Jeff
> >
> > jeff.pr2 (at) added-insight (dot) net
> >