[R] CART analysis

[Bernardo Rangel Tura]

On Sun, 2007-12-09 at 21:24 +0100, anna.paganoni at polimi.it wrote:
> I would like to know if is it possible implemet a partitioning tree  
> using a function like rpart, or mvpart, and with formula a glm-object  
> (as a logistic models) or a robust linear regression (as least sum of  
> absolute errors).
> In this case, the appropriate "method" to use is "mrt"? Or another one?
> Thanks,
> Anna Maria Paganoni

Well, 


I think de package tree solve your problem. If your response variable
don't use logistic approach if your variable is numeric use standard
formula

Any doubt read :
http://cran.r-project.org/doc/packages/tree.pdf


and mail me 

-- 
Bernardo Rangel Tura, M.D,MPH,Ph.D
National Institute of Cardiology
Brazil