Thanks to visit codestin.com
Credit goes to github.com

Skip to content

ENH preserves dtypes for GenericUnivariateSelect class - #22370

Merged
jeremiedbb merged 18 commits into
scikit-learn:mainfrom
Micky774:feature_dtype
Feb 14, 2022
Merged

ENH preserves dtypes for GenericUnivariateSelect class#22370
jeremiedbb merged 18 commits into
scikit-learn:mainfrom
Micky774:feature_dtype

Conversation

@Micky774

@Micky774 Micky774 commented Feb 4, 2022

Copy link
Copy Markdown
Contributor

Reference Issues/PRs
Addresses #11000
Picks up stalled PR #18482

What does this implement/fix? Explain your changes.
PR #18482: Preserves dtypes for GenericUnivariateSelect class.
This PR: Picks up stall and merges main

Any other comments?

@Micky774

Micky774 commented Feb 4, 2022 •
edited
Loading

Copy link
Copy Markdown
Contributor Author

@rth @cmarmo

Re: "...we need to investigate why test_estimators[GenericUnivariateSelect()-check_dtype_object(strict_mode=True)] is failing"
After looking into it for a bit, it seems that the GUS performs checks on its X, y data explicitly through the _check_params method of the chosen selector and implicitly in the passed score function. In this case, functions like chi2 successfully check against the object case, but evidently f_regression -- or more precisely, r_regression -- doesn't.

Given that r_regression makes sense when called with a potentially-numeric object-type y, I think it would be reasonable to just set y_numeric=True in

scikit-learn/sklearn/feature_selection/_univariate_selection.py

Line 295 in 9f85c9d

X, y = check_X_y(X, y, accept_sparse=["csr", "csc", "coo"], dtype=np.float64)

thomasjpfan
thomasjpfan reviewed Feb 4, 2022
View reviewed changes

@thomasjpfan thomasjpfan left a comment

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks for the PR!

Comment thread sklearn/utils/estimator_checks.py Outdated
Comment thread sklearn/feature_selection/_base.py
Comment thread doc/whats_new/v1.1.rst Outdated
thomasjpfan
thomasjpfan reviewed Feb 7, 2022
View reviewed changes

@thomasjpfan thomasjpfan left a comment

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Looking good.

Comment thread sklearn/feature_selection/tests/test_feature_select.py Outdated
Comment thread sklearn/utils/estimator_checks.py Outdated
Comment thread doc/whats_new/v1.1.rst Outdated
thomasjpfan
thomasjpfan reviewed Feb 11, 2022
View reviewed changes
Comment thread sklearn/utils/estimator_checks.py Outdated
thomasjpfan
thomasjpfan approved these changes Feb 12, 2022
View reviewed changes

@thomasjpfan thomasjpfan left a comment

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks for the update! LGTM

@thomasjpfan thomasjpfan added the Quick Review For PRs that are quick to review label Feb 12, 2022
jeremiedbb
jeremiedbb approved these changes Feb 14, 2022
View reviewed changes

@jeremiedbb jeremiedbb left a comment

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM. Thanks @Micky774 !

@jeremiedbb
jeremiedbb merged commit e821a9e into scikit-learn:main Feb 14, 2022
@jeremiedbb jeremiedbb mentioned this pull request Feb 14, 2022
Closed
thomasjpfan added a commit to thomasjpfan/scikit-learn that referenced this pull request Mar 1, 2022
@Micky774 Thierry Gameiro
@titigmr @thomasjpfan
ENH preserves dtypes for GenericUnivariateSelect class (scikit-learn#…
7c2c2c8 
…22370)

Co-authored-by: Thierry Gameiro <[email protected]>
Co-authored-by: titigmr <[email protected]>
Co-authored-by: Thomas J. Fan <[email protected]>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@thomasjpfan thomasjpfan thomasjpfan approved these changes
@jeremiedbb jeremiedbb jeremiedbb approved these changes

Assignees

No one assigned

Labels

module:feature_selection module:utils Quick Review For PRs that are quick to review

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

4 participants

@Micky774 @thomasjpfan @jeremiedbb @titigmr