Add functions to conduct k-fold cross validations

[AdamCSmithCWS]

Recent work has shown that WAIC performs very poorly with BBS data. Full leave one out CV is not feasible. However, a k-fold (e.g., k = 10) cross validation could be done within a reasonable amount of time.
We will need to ensure that the random sampling is done in a stratified way, such that each fold retains enough data to estimate all of the stratum-specific parameters.