Root Mean Square Error in Generalized Regression and assessment methodsWhy doesn't RMSE show up in some assessment methods like Leave-One-Out and K-Fold? Any suggestions for comparing these models with holdback? The entire "Measure" box dissapears and RMSE, etc. along with it (for example just stick with LASSO and changet the assessment or "validation" method). But... we see a -Loglikelihood being graphed next to the solution path. Why don't we get the model fit stastitcs written out?
Re: Root Mean Square Error in Generalized Regression and assessment methodsNot sure why the RMSE doesn't show up...but to perhaps help answer your question, "Any suggestions for comparing these models with holdback?" Can we use the Scale term? My understanding of this term in the Table of Parameter estimates is this is the moral equivalent of RMSE in an OLS context? At least that's how I interpret the notes @di_michelson handed out at the dress rehearsal for "Finding Important Predictors" during the Mid Year event...see page 4-15 of the course notes.
