[R] About separate train and test data

javed khan j@vedbtk111 @end|ng |rom gm@||@com
Fri Nov 8 18:56:02 CET 2019


For instance, we have separate train and test data files (not want to do k
fold), so we will not use the function trainControl? In that case if we
have to tune the parameters, do we need to specify search =grid in the
train function?

My second question is how we can measure MCC classification measure? Is it
same like metric=Roc in the train function?

	[[alternative HTML version deleted]]

More information about the R-help mailing list