Integer. Number of splits for the V-fold cross-validation step. The default is 10. In most cases, between 10 and 20 splits works well.

Logical. Should the data splits be stratified by a binary response? Attempts to maintain the same ratio in each training and validation sample.

Logical. Should the rows of X be shuffled before creating the splits.

A List. Use this to pass pre-specified rows for the sample splits. The length of the list should be V and each entry in the list should contain a vector with the row numbers of the corresponding validation sample.