R help - Oct 2008 - Problem with glm.nb estimation

Dear All,

I've been using already for a year glm.nb() from the MASS package.
But today, R gave me an error message when estimating one of my usual 
models:

> depEsf.nb <- glm.nb(depE ~ manuf00E + corps00E + lngdp00E + lngdp00sqE +
> lnpop00E + indshE + scishE + mechshE + elecshE + chemshE + drugshE + 
> urban_dummyE + aggl_dummyE
+ + eE1 + eE2 + eE3 + eE4 + eE5 + eE6 + eE7 + eE8 + eE9 + eE10 + eE11 + eE12 
+ eE13 + eE14 + eE15 + eE16 + eE17, maxit = 100000)
Warning messages:
1: In sqrt(1/i) : NaNs produced
2: In sqrt(1/i) : NaNs produced


What does it mean to receive such a message? And, most of all, can I fix 
this problem?
The only problem I could think of is that, for this particular case, I do 
not have lots of degrees of freedom (N = 67), and I use 30 variables.
Even if I start from a basic model (no eE... variables) and try to build the 
model up by step(), the stepwise procedure fails at the first step (I know 
step() is not made for the glm.nb() function, but it usually works just fine 
for me).

Thanks in advance to everyone who might drop his/her two cents on what the 
problem is.

Sincerely,
Roberto

********************
Roberto Patuelli, Ph.D.
Post-doc researcher
Institute for Economic Research (IRE)
University of Lugano (USI)
via Maderno 24, CP 4361
CH-6904 Lugano
Switzerland
Phone: +41-(0)58-666-4166
Email: roberto.patuelli at lu.unisi.ch

I see no error message here -- a *warning* is not an error.

Please give a fully reproducible example with the actual output containing 
an error message.

On Fri, 3 Oct 2008, Roberto Patuelli wrote:
> Dear All,
>
> I've been using already for a year glm.nb() from the MASS package.
> But today, R gave me an error message when estimating one of my usual
models:
>
>
>> depEsf.nb <- glm.nb(depE ~ manuf00E + corps00E + lngdp00E +
lngdp00sqE +
>> lnpop00E + indshE + scishE + mechshE + elecshE + chemshE + drugshE + 
>> urban_dummyE + aggl_dummyE
> + + eE1 + eE2 + eE3 + eE4 + eE5 + eE6 + eE7 + eE8 + eE9 + eE10 + eE11 +
eE12
> + eE13 + eE14 + eE15 + eE16 + eE17, maxit = 100000)
> Warning messages:
> 1: In sqrt(1/i) : NaNs produced
> 2: In sqrt(1/i) : NaNs produced
>
>
> What does it mean to receive such a message? And, most of all, can I fix
this
> problem?
> The only problem I could think of is that, for this particular case, I do
not
> have lots of degrees of freedom (N = 67), and I use 30 variables.
> Even if I start from a basic model (no eE... variables) and try to build
the
> model up by step(), the stepwise procedure fails at the first step (I know 
> step() is not made for the glm.nb() function, but it usually works just
fine
> for me).
>
> Thanks in advance to everyone who might drop his/her two cents on what the 
> problem is.
>
> Sincerely,
> Roberto
>
> ********************
> Roberto Patuelli, Ph.D.
> Post-doc researcher
> Institute for Economic Research (IRE)
> University of Lugano (USI)
> via Maderno 24, CP 4361
> CH-6904 Lugano
> Switzerland
> Phone: +41-(0)58-666-4166
> Email: roberto.patuelli at lu.unisi.ch
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>
-- 
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