Generates a list of folds for a variety of cross-validation schemes.

make_folds(
  n = NULL,
  fold_fun = folds_vfold,
  cluster_ids = NULL,
  strata_ids = NULL,
  ...
)

Arguments

n

- either an integer indicating the number of observations to cross-validate over, or an object from which to guess the number of observations; can also be computed from strata_ids or cluster_ids.

fold_fun

- A function indicating the cross-validation scheme to use. See fold_funs for a list of possibilities.

cluster_ids

- a vector of cluster ids. Clusters are treated as a unit -- that is, all observations within a cluster are placed in either the training or validation set.

strata_ids

- a vector of strata ids. Strata are balanced: insofar as possible the distribution in the sample should be the same as the distribution in the training and validation sets.

...

other arguments to be passed to fold_fun.

Value

A list of folds objects. Each fold consists of a list with a training index vector, a validation index vector, and a fold_index (its order in the list of folds).

See also

Other fold generation functions: fold_from_foldvec(), fold_funs, folds2foldvec(), make_repeated_folds()