Phil has one equation typo (to hand me). He says the document treats the Evidence as the right thing to do and cross validation as a competitor. It should be more balanced. It should also explain the problem with the evidence and uninformative priors (the answer is just settable with the priors). Cross validation is not well explained for those who haven't seen it previously.
Phil has one equation typo (to hand me). He says the document treats the Evidence as the right thing to do and cross validation as a competitor. It should be more balanced. It should also explain the problem with the evidence and uninformative priors (the answer is just settable with the priors). Cross validation is not well explained for those who haven't seen it previously.