[R] randomForest: sampling with replacement?

Max Kuhn mxkuhn at gmail.com
Mon May 10 14:51:28 CEST 2010

>> Would you please confirm that it is a bootstrap sample with
>> replacement?

Someone should note that the definition of a bootstrap sample is a
sample with replacement (usually of size n).

I've read quite a few papers where they claim to be using the
bootstrap. Upon further review (sometimes to the code) they often are
doing permutation testing-type randomization. This is mostly in
biological journals.

There are also at least paper that describe the "out of box" random
forests error rate...


More information about the R-help mailing list