Create Cross-Fitted Learner — make_cross

This function takes a learner and returns a cross-fitted version of it. Cross-fitting involves fitting the learner to different subsets of the data while using the complementary subsets for validation, to provide a more robust estimate of model performance.

Usage

make_cross_fitted(
  learner,
  calibrate = FALSE,
  cross_validate = inherits(learner, "Stack")
)

Arguments

learner: The learner to be cross-fitted. This can be a single learner or a stacked learner.
calibrate: Currently not used. Logical indicating whether to calibrate the learner (default is FALSE).
cross_validate: Logical indicating whether to perform cross-validation (default is TRUE if learner is a stacked learner).

Value

A cross-fitted version of the input learner.