R devel - Aug 2010 - Decision tree and factor variables

Hello,

I'm building a decision tree in R with the rpart package. Modeling is
fine. But when it comes to scoring, I have the following issue:

factor 'cust_language' has new level(s) OT

I think this comes from the fact that when learning, the DT doesn't
see all the possible value of the factor variable cust_language. When
scoring, new values comes and I get this error. However, it should not
be a problem to have new values for a factor variable when scoring
with decision tree.

Any idea on how I should handle the problem?
Thanks. 
-- 
View this message in context:
http://r.789695.n4.nabble.com/Decision-tree-and-factor-variables-tp2339234p2339234.html
Sent from the R devel mailing list archive at Nabble.com.

On Thu, 2010-08-26 at 00:06 -0700, clusty wrote:
> Hello,
> 
> I'm building a decision tree in R with the rpart package. Modeling is
> fine. But when it comes to scoring, I have the following issue:
> 
> factor 'cust_language' has new level(s) OT
> 
> I think this comes from the fact that when learning, the DT doesn't
> see all the possible value of the factor variable cust_language. When
> scoring, new values comes and I get this error. However, it should not
> be a problem to have new values for a factor variable when scoring
> with decision tree.
> 
> Any idea on how I should handle the problem?
> Thanks. 
Wrong list. R-Devel is for discussion pertaining to development of and
with R. You need R-Help.

G

-- 
%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%
 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
%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%