1

I'm currently trying to understand the Bootstrap method. I have a rather simple question. Say I have my original set S of size N and wish to create B bootstrapped samples of size R. Hastie et al. states in 'The Elements of Statistical Learning' ,p. 249, that

"The basic idea is to randomly draw datasets with replacement from the training data, each sample the same size as the original training set."

However, looking around on various online sources it appears that the sizes of the bootstrapped samples is varying in size and can be arbitrarily chosen, with a discussion to be held on what the proper size of the bootstrapped samples.

So my question is if the size of the bootstrapped samples the same as the size of the original dataset?

Kindt
  • 11
  • @einar Hmm not really, since he's doing the bootstraping samples as the same size as the original samples, but I still think it's possible to create bootstrap samples of lesser size than the original... – Kindt Mar 23 '22 at 10:28
  • 1
    True. How about https://stats.stackexchange.com/questions/220013/can-we-use-bootstrap-samples-that-are-smaller-than-original-sample/220866 and https://stats.stackexchange.com/questions/33948/subsample-bootstrapping – einar Mar 23 '22 at 12:30
  • I agree with @einar 's second comment, these last two threads are answering the OP's question (as I understand it at least). – usεr11852 Mar 23 '22 at 13:48

0 Answers0