ENH Adds n_features_in_ checking in cross_decomposition #18741
Conversation
|
|
|
conflict... |
|
error seems related |
|
tests still failing |
sklearn/utils/estimator_checks.py
Outdated
| if is_regressor(estimator): | ||
| y = rng.normal(size=n_samples) | ||
| else: | ||
| y = rng.randint(low=0, high=2, size=n_samples) | ||
| y = _enforce_estimator_tags_y(estimator, y) | ||
|
|
||
| if name in CROSS_DECOMPOSITION: |
There was a problem hiding this comment.
could you please add a comment as why this is needed for future reference?
There was a problem hiding this comment.
is it making sense to have this in _enforce_estimator_tags_y with proper tags?
There was a problem hiding this comment.
As a quick fix, we acn put it into _enforce_estimator_tags_y first without a tag. We do this name in CROSS_DECOMPOSITION "hack" a bunch in estimator_checks, it would be worth it to centralize.
An alternative would be to n_components=1 in _set_checking_parameters for the cross_decomposition. I went with this option for this PR.
|
Thanks @thomasjpfan |
Continues #18514