LFD Book Forum (http://book.caltech.edu/bookforum/index.php)
-   Homework 7 (http://book.caltech.edu/bookforum/forumdisplay.php?f=136)
-   -   Interaction of cross validation with model selection (http://book.caltech.edu/bookforum/showthread.php?t=4012)

 hemphill 02-20-2013 12:09 PM

Interaction of cross validation with model selection

With 10-fold cross validation, we do 10 training runs. How would you recommend we do model selection? If we have M models, do we do 10*M training runs? Or do we do model selection with simple validation, then use cross validation for an error estimate? There seem to be a lot of possibilities. If we wish to use validation to choose an "early stopping" parameter, we could estimate the value of this parameter in each of the cross validation runs, then use the average when training with the full data set. Is this OK?

 yaser 02-20-2013 01:57 PM

Re: Interaction of cross validation with model selection

Quote:
 Originally Posted by hemphill (Post 9451) With 10-fold cross validation, we do 10 training runs. How would you recommend we do model selection? If we have M models, do we do 10*M training runs? Or do we do model selection with simple validation, then use cross validation for an error estimate? There seem to be a lot of possibilities. If we wish to use validation to choose an "early stopping" parameter, we could estimate the value of this parameter in each of the cross validation runs, then use the average when training with the full data set. Is this OK?
10-fold cross validation for selecting a model among M models will indeed take 10*M training sessions. If you want to choose the value of a parameter, you save all ten cross-validation errors for the different values of the parameter, average these errors for each parameter value, then pick the parameter value that has the smallest average error.

 All times are GMT -7. The time now is 04:16 AM.