[R] rpart$where and predict.rpart

[Prof Brian Ripley]

On Tue, 22 Jul 2008, Chua Siang Li wrote:

>
>   Hello there.  I have fitted a rpart model.
>   > rpartModel <- rpart(y~., data=data.frame(y=y,x=x),method="class", ....)
>   and  can  use  rpart$where  to  find  out the terminal nodes that each
>   observations belongs.
>   Now, I have a set of new data and used predict.rpart which seems to give
>   only the predicted value with no information similar to rpart$where.
>   May I know how to find out the terminal nodes that each new observations
>   belongs to?

You need to read the code to answer questions about internals.  A very 
quick glance at predict.rpart would have told you that

         where <- pred.rpart(object, rpart.matrix(newdata))

is what you are looking for.  (Note that pred.rpart like many other 
internal functions is not exported from the name space.)

>   Many thanks.
>   ----
>   Chua Siang Li
>   Consultant - Operations Research
>   Acceval Pte Ltd
>   Tel: 6297 8740
>   Email: siang.li.chua at acceval-intl.com
>   Website: www.acceval-intl.com

-- 
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