[R] RWeka evaluate classifier on test set

Uwe Ligges ligges at statistik.tu-dortmund.de
Fri Jun 19 13:26:49 CEST 2009



Huihua Lu wrote:
> Hi everyone,
> 
> I have a test set with more than 1000 cases, when I use 
> evaluate_Weka_classifier(RWeka)to evaluate my classifier on this test 
> set, the output shows me the result of only 83 cases. I do have 
> missing values in predictors, so I tried na.acton=na.pass, but it 
> dosen't help. Now I confused, why RWeka ignore so many cases in my 
> test set? Is there any setting I didn't notice in 
> evaluate_Weka_classifier(RWeka)? Anyone can help? Thanks in advance!


Well, without knowing Weka: You need a classification method that 
supports NAs. Hence typically tree based methods that make use of 
surrogate variables.

Best,
Uwe Ligges




> lu
> 
> ______________________________________________
> R-help at r-project.org mailing list
> 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.




More information about the R-help mailing list