[R] #library(party) - Compare predicted results for ctree

[Achim Zeileis]

On Mon, 16 Feb 2015, Rodica Coderie via R-help wrote:

> Hello,
>
> I've created a ctree model called fit using 15 input variables for a factor predicted variable Response (YES/NO).
> When I run the following :
> table(predict(fit2), training_data$response)
> I get the following result:
>
>     NO   YES
> NO  48694   480
> YES     0     0
>
> It appears that the NO responses are predicted with 100% accuracy and 
> the YES response are predicted with 0% accuracy.
>
> Why is this happening? It's because of my data or it's something in 
> ctree algorithm?

Your data has less than 1% of YES observations and I would guess that the 
tree cannot separate these in a way such that majority voting gives a YES 
prediction. You might consider a different cutoff (other than 50%) or 
downsampling the NO observations.

> Thanks!
> Rodica
>
> ______________________________________________
> R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>