ertk.train.get_cv_splitter
- ertk.train.get_cv_splitter(group: bool, k: int, test_size: float = 0.2, shuffle: bool = False, random_state: int | None = None) BaseCrossValidator
Gets an appropriate cross-validation splitter for the given number of folds and groups, or a single random split.
- Parameters:
- group: bool
Whether to split over pre-defined groups of instances.
- k: int
If k > 1 then do k-fold CV. If k == 1 then do one random split. If k = -1 then do leave-one-out. If k == 0 then use the whole train set as validation split.
- test_size: float
The size of the test set when k == 1 (one random split).
- shuffle: bool
Whether to shuffle when using k-fold for k > 1.
- random_state: int, optional
The random state to set for splitters with shuffling behaviour.
- Returns:
- splitter: BaseCrossValidator
Cross-validation splitter that has
split()andget_n_splits()methods.