Since bagging seems to reduce variance without increasing the size of the hypothesis set (I think), is it fair to say that it does not increase the bias? Therefore, in terms of out-of-sample error, it is always a good thing?

I know that there may be costs associated with interpretability and computational cost from bagging, but just considering the model's performance out of sample, are there any downsides?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.