skada.model_selection.StratifiedDomainShuffleSplit

class skada.model_selection.StratifiedDomainShuffleSplit(n_splits=10, *, test_size=None, train_size=None, random_state=None)

Stratified-Domain-Shuffle-Split cross-validator.

This cross-validation object returns stratified randomized folds. The folds are made by preserving the percentage of samples for each class and for each sample domain.

Parameters:

n_splitsint, default=10

Number of folds. Must be at least 2.

Examples

>>> import numpy as np
>>> from skada.model_selection import StratifiedDomainShuffleSplit
>>> X = np.ones((10, 2))
>>> y = np.array([-1, 0, 1, -1, 0, 1, -1, 0, 1, -1])
>>> sample_domain = np.array([-2, 1, 1, -2, 1, 1, -2, 1, 1, -2])
>>> da_shufflesplit = StratifiedDomainShuffleSplit(n_splits=2,
... random_state=0, test_size=0.5)
>>> da_shufflesplit.get_n_splits(X, y, sample_domain)
2
>>> print(da_shufflesplit)
StratifiedDomainShuffleSplit(n_splits=2, random_state=0,
    test_size=0.5, train_size=None)
>>> for i, (train_index, test_index) in enumerate(
...    da_shufflesplit.split(X, y, sample_domain)
... ):
...     print(f"Fold {i}:")
...     print(f"  Train: index={train_index}, "
...     f'''group={[[b.item(), a.item()]
...     for a, b in zip(y[train_index], sample_domain[train_index])
...     ]}''')
...     print(f"  Test: index={test_index}, "
...     f'''group={[[b.item(), a.item()]
...     for a, b in zip(y[test_index], sample_domain[test_index])
...     ]}''')
Fold 0:
    Train: index=[0 6 1 8 2], group=[[-2, -1], [-2, -1], [1, 0], [1, 1], [1, 1]]
    Test:  index=[4 9 7 5 3], group=[[1, 0], [-2, -1], [1, 0], [1, 1], [-2, -1]]
Fold 1:
    Train: index=[1 2 8 0 3], group=[[1, 0], [1, 1], [1, 1], [-2, -1], [-2, -1]]
    Test:  index=[7 5 9 4 6], group=[[1, 0], [1, 1], [-2, -1], [1, 0], [-2, -1]]

set_split_request(*, sample_domain: bool | None | str = '$UNCHANGED$') → StratifiedDomainShuffleSplit

Request metadata passed to the split method.

Note that this method is only relevant if enable_metadata_routing=True (see sklearn.set_config()). Please see User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to split if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to split.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Note

This method is only relevant if this estimator is used as a sub-estimator of a meta-estimator, e.g. used inside a Pipeline. Otherwise it has no effect.

Parameters:

sample_domainstr, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED

Metadata routing for sample_domain parameter in split.

Returns:

selfobject

The updated object.

split(X, y, sample_domain=None)

XXX: Docstring here

Examples using skada.model_selection.StratifiedDomainShuffleSplit

Visualizing cross-validation behavior in skada

Visualizing cross-validation behavior in skada