[R] CART and the `tree' contrib package

[Prof Brian D Ripley]

On Mon, 20 Mar 2000, Faheem Mitha wrote:

> Dear R people,
> 
> I was recently reading the book `Classification and Regression Trees' by
> Breiman. This book talks about the CART program. Both Splus and R have
> implementations of this. However, the book talks about the possibility of

No, they don't.  CART is a trademark. Both S and R have tree-based
modelling (tree) and something similar to CART (rpart).

> extending the existing `standard' set of questions (for continuous
> variables, these are of the form X < c where X is the variable, c some
> const) to questions of the form (for continuous variables) 
> sum over i a_i X_i < c where X_i are the cont. variables, a_i and c
> consts.
> 
> However, this extended set is not implemented in either the R `tree'
> package nor the Splus one, to the best of my knowledge. (I believe there
> is some original CART package, but I don't know about that). Anyway, I was
> wondering why not. Does anyone know?

Because it is very much harder to optimize over such splits, and the CART
program is not very good at this.  There exists many other variations on th
this theme: see my PRNN book.

-- 
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 272860 (secr)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._