[R] rpart with only two small groups

[Gavin Simpson]

On Tue, 2008-09-16 at 10:47 -0700, Birgitle wrote:
> Hello R-User!
> 
> I try to do the following:
> 
> New<-iris[c(1:7,90:97),1:5]
> New.rpart<-rpart(Species~., data=New, method="class")
> 
> New.rpart
> n= 15 
> 
> node), split, n, loss, yval, (yprob)
>       * denotes terminal node
> 
> 1) root 15 7 versicolor (0.4666667 0.5333333) *
> 
> Does it mean it is not possible to find a variable that splits my groups or
> are my groups just to small?
> Is there a way to find out the variable that splits my groups best?

Yes, as always, read the help file ;-)

?rpart leads you to ?rpart.control, which has argument minsplit with
default 20, which is the minimum number of observations in a node for a
split to be attempted. As 15 is less than 20, the root node could not be
split.

E.g.:
> New<-iris[c(1:7,90:97),1:5]
> New.rpart<-rpart(Species~., data=New, method="class", control =
rpart.control(minsplit = 2))
> New.rpart
n= 15 

node), split, n, loss, yval, (yprob)
      * denotes terminal node

1) root 15 7 versicolor (0.4666667 0.5333333)  
  2) Petal.Length< 2.5 7 0 setosa (1.0000000 0.0000000) *
  3) Petal.Length>=2.5 8 0 versicolor (0.0000000 1.0000000) *

HTH

G

> 
> Thanks in advance for help.
> 
> B.
> 
> -----
> The art of living is more like wrestling than dancing.
> (Marcus Aurelius)
-- 
%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%
 Dr. Gavin Simpson             [t] +44 (0)20 7679 0522
 ECRC, UCL Geography,          [f] +44 (0)20 7679 0565
 Pearson Building,             [e] gavin.simpsonATNOSPAMucl.ac.uk
 Gower Street, London          [w] http://www.ucl.ac.uk/~ucfagls/
 UK. WC1E 6BT.                 [w] http://www.freshwaters.org.uk
%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%