Concept - Cross Validation
On the slides (s.247): cross validation can help decide which things (popcorn, soda pop and/or candy) help make the best classifications, i.e. which features to keep in the Concept - Gaussian Naive Bayes model. Only this use is stated in this chunk.
Added context
Generally: split data into k folds, train on k-1, evaluate on the held-out fold, average the scores, and compare feature subsets/models by that score.