Lots of supervised learning theory is motivated using the IID assumption. Do most of these methods apply equally well if data is only exchangeable, and not IID? Can you provide an example where this is not the case, and explain why?
Are there examples of ML or stats approaches that are valid for IID data, but not exchangeable data?
Asked
Active
Viewed 40 times
3
-
Some relevant posts: https://stats.stackexchange.com/questions/344794/exchangeability-and-iid-random-variables, https://stats.stackexchange.com/questions/344794/exchangeability-and-iid-random-variables/344830#344830 – kjetil b halvorsen Sep 01 '22 at 23:20