R help - Oct 2005 - stepwise on generlized linear model

Hi-

Can anybody give me a quick answer that whether step() can be directly 
applied to a glm model including both continuous variables and categorical 
variables (by factor())? Maybe my data is not suitable, I applied step() on 
my glm object and did not get any reduction (i.e., returned full model).

Can gl1ce() in lasso2 deal with glm model with both type variables?

Thanks,
WC

On Sun, 16 Oct 2005, Weijie Cai wrote:
> Can anybody give me a quick answer that whether step() can be directly
> applied to a glm model including both continuous variables and categorical
> variables (by factor())? Maybe my data is not suitable, I applied step() on
> my glm object and did not get any reduction (i.e., returned full model).
Yes, it can.  There are examples in e.g. MASS4.

-- 
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