[R] gee, error=104

[Prof Brian Ripley]

On Wed, 10 Oct 2007, Michel Chavance wrote:

> I am using gee to fit a log-linear marginal model to grouped data. I
> get sensible results unfortunately followed by this warning
>
> Warning message:
> Note: Cgee had an error (code= 104 ).  Results suspect. in:
> gee(formula = event ~ exgr + agegr + indiv + offset(loginterval),
>
> Does anyone knows what is a code=104 error ?

Anyone who reads the source code does.  Look in gee/src/ugee.h.

-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595