Hello!
I'm Taeho, a graduate student in South Korea.
In order to get .632+ bootstrap error using random forest, I have tried to use
'ipred' package; more specifically the function 'errorest' has been used.
Following the guidelines, I made a simple command line like below:
error<-errorest(class ~ ., data=data, model=randomForest, estimator =
"632plus")$err
however, I got an error message saying that:
randomForest.default(m, y, ...) :
Can't have empty classes in y.
The data I used includes three classes and each of the classes has 5, 2, and 4
samples, respectively. Other than this fact about the data, nothing is
different from the example 'iris' data. (By the way, I could get a simple OOB
error for this data without any problem just using 'randomForest' function
only.)
What do you think my problem is?
I think the problem is that the number of samples is too small to run
'errorest' for getting .632 plus bootstrap error. Is this correct? If so, what
would you recommend to fix this problem?
Thank you for reading this and for your kind answer in advance.
------------------------------------------------------------------------
새로운 기부 문화의 씨앗, 해피빈
http://happybean.naver.com
______________________________________________
[email protected] 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.
