5.5 k-fold Cross-Validation

• This is an alternative to LOOCV which involves dividing the data set into \(k\) groups (folds) of approximately equal size.
• The percent of the data set that is in the validation set can be thought of as \(1/k\).
• E.g., for \(k=5\) groups, 20% of the data would be withheld for testing.