3
$\begingroup$

I get it, cross validation works. I'm wondering if there is an literature out there giving any theoretical justification for cross validation. My thought is that there should be, at least, something like the central limit theorem. Any pointers to literature, or comments are most welcome. Thanks.

$\endgroup$
3
$\begingroup$

A quick search on Google Scholar produced quite a few papers. Here are two of the most highly cited ones:

A study of cross-validation and bootstrap for accuracy estimation and model selection (PDF)

A leisurely look at the bootstrap, the jackknife, and cross-validation (PDF)

$\endgroup$
  • $\begingroup$ @ aplassard. The first article is very interesting, but totally empirical. The second, is wonderful. I think that pretty much anytime I read something by Efron. For the record, section 8 discusses cv, and in section 9 there are comments explaining the relationship between cv and bootstrap that are really interesting. $\endgroup$ – meh Apr 24 '14 at 12:59

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.