R help - Jan 2003 - Starting values for glm fits

I'm fitting some small data sets using a binomial glm with complementary 
log-log link. In some ill-conditioned cases I am getting convergence 
failures. I know how to adjust maxit and epsilon, but that doesn't seem 
to help.

In fact I know some very good starting values for my fits, but I can't 
see how to get them in to glm(). How may I do this?

-- 
Dr Murray Jorgensen      http://www.stats.waikato.ac.nz/Staff/maj.html
Department of Statistics, University of Waikato, Hamilton, New Zealand
Email: maj at waikato.ac.nz                                Fax 7 838 4155
Phone  +64 7 838 4773 wk    +64 7 849 6486 home    Mobile 021 1395 862

>>>>> "Murray" == Murray Jorgensen <maj at
stats.waikato.ac.nz>
>>>>>     on Tue, 21 Jan 2003 13:47:16 +1300 writes:
    Murray> I'm fitting some small data sets using a binomial
    Murray> glm with complementary log-log link. In some
    Murray> ill-conditioned cases I am getting convergence
    Murray> failures. I know how to adjust maxit and epsilon,
    Murray> but that doesn't seem to help.

    Murray> In fact I know some very good starting values for my
    Murray> fits, but I can't see how to get them in to
    Murray> glm(). How may I do this?

?glm aka help(glm) -- tells a quite a bit about this:

Overly terse summary:

   glm(*, start = ..)

or for more possibilities (and less ease of use)
   glm.fit(*, etastart = .)
or
   glm.fit(*, start = .)
or
   glm.fit(*, mustart = .)

-----
Actually I think, glm() could easily be enhanced to also accept
`etastart' and `mustart' since all of these "start" arguments
would just be passed to glm.fit() ...

Martin Maechler <maechler at stat.math.ethz.ch>
http://stat.ethz.ch/~maechler/
Seminar fuer Statistik, ETH-Zentrum  LEO C16	Leonhardstr. 27
ETH (Federal Inst. Technology)	8092 Zurich	SWITZERLAND
phone: x-41-1-632-3408		fax: ...-1228			<><