8000 FIX Allow input validation by pass in MLPClassifier by betatim · Pull Request #24873 · scikit-learn/scikit-learn · GitHub
[go: up one dir, main page]
Skip to content
/ < 8000 a data-pjax="#repo-content-pjax-container" data-turbo-frame="repo-content-turbo-frame" href="/scikit-learn/scikit-learn">scikit-learn Public

FIX Allow input validation by pass in MLPClassifier #24873

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Merged
merged 10 commits into from
Jan 4, 2023
Merged

FIX Allow input validation by pass in MLPClassifier #24873

merged 10 commits into from
Jan 4, 2023

Conversation

betatim
Copy link
Member
@betatim betatim commented Nov 9, 2022 •
edited
Loading

With arrays created inside the estimator we do not need to perform validation again. This change allows the estimator to compute the score during fitting, without having to provide column names for the validation dataset.

This solves a problem where fitting data with feature names generates warnings. The reason for the warnings is that the input validation strips the column names from the input data, which is then split into a train and validation set. When the validation set is passed to score, it performs input validation and raises a warning because the input data doesn't have feature names.

Reference Issues/PRs

closes #24846

What does this implement/fix? Explain your changes.

This adds a private score method which is used during fitting with early stopping. The private score method uses a _predict method that does not perform input data validation. This is Ok because the validation dataset is derived from a train_test_split of "input validated data".

Alternatives

We could turn X_val into a dataframe when it is created by using self.feature_names_in_. Feels awkward.

We could delay setting feature_names_in_ until the fitting is complete, would require reworking the validation functions.

What do people think?

@betatim
Allow input validation by pass
1a4e0d9 
With arrays created inside the estimator we do not need to perform
validation again. This change allows the stimator to compute the score
during fitting, without having to provide column names for the
validation dataset.
@betatim
Add validation bypass for MLPRegressor
ed9291e
@cmarmo
Copy link
Contributor
cmarmo commented Nov 11, 2022

Hi @betatim, thanks for the pull request ( :) ).
Pandas is still a soft dependency for scikit-learn tests.
You might want to use pytest.importorskip as in

scikit-learn/sklearn/datasets/tests/test_20news.py

Line 104 in e947074

pd = pytest.importorskip("pandas")
to fix the checks.

@betatim
Copy link
Member Author
betatim commented Nov 14, 2022

TIL about importorskip!

@betatim
Copy link
Member Author
betatim commented Nov 14, 2022

This is ready for review. No Changelog entry because I don't think this is needed for this change.

@cmarmo cmarmo added No Changelog Needed Validation related to input validation labels Nov 14, 2022
thomasjpfan
thomasjpfan reviewed Nov 23, 2022
View reviewed changes
Copy link
Member
@thomasjpfan thomasjpfan left a comment

Choose a reason for hiding this comment

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

Thank you for the PR @betatim !

@betatim
Copy link
Member Author
betatim commented Nov 24, 2022

All comments addressed.

thomasjpfan
thomasjpfan reviewed Dec 1, 2022
View reviewed changes
Copy link
Member
@thomasjpfan thomasjpfan left a comment

Choose a reason for hiding this comment

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

I think this is a big enough change to have a change log entry as a bug fix.

@betatim
Copy link
Member Author
betatim commented Dec 2, 2022

Added what's new entry, targeting v1.3. Not sure about how/when things still get added to the upcoming v1.2 but I guess we can do that in addition?

@thomasjpfan thomasjpfan changed the title ENH Allow input validation by pass in MLPClassifier FIX Allow input validation by pass in MLPClassifier Dec 2, 2022
thomasjpfan
thomasjpfan approved these changes Dec 5, 2022
View reviewed changes
Copy link
Member
@thomasjpfan thomasjpfan left a comment

Choose a reason for hiding this comment

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

LGTM!

thomasjpfan
thomasjpfan reviewed Dec 5, 2022
View reviewed changes
@betatim betatim added the Waiting for Second Reviewer First reviewer is done, need a second one! label Dec 6, 2022
@glemaitre glemaitre self-requested a review December 28, 2022 11:01
@glemaitre glemaitre removed the Waiting for Second Reviewer First reviewer is done, need a second one! label Dec 28, 2022
@betatim
Copy link
Member Author
betatim commented Jan 4, 2023

Thanks for the review @glemaitre. I've implemented your changes.

@glemaitre glemaitre merged commit 97fe999 into scikit-learn:main Jan 4, 2023
@glemaitre
Copy link
Member

LGTM. Thanks @betatim

@betatim betatim deleted the feature-names-mlp branch January 4, 2023 10:02
jpangas pushed a commit to jpangas/scikit-learn that referenced this pull request Jan 4, 2023
@betatim @jpangas
FIX Allow input validation by pass in MLPClassifier (scikit-learn#24873)
5af83b6
jjerphan pushed a commit to jjerphan/scikit-learn that referenced this pull request Jan 20, 2023
@betatim @jjerphan
FIX Allow input validation by pass in MLPClassifier (scikit-learn#24873)
66fd257
jjerphan pushed a commit to jjerphan/scikit-learn that referenced this pull request Jan 20, 2023
@betatim @jjerphan
FIX Allow input validation by pass in MLPClassifier (scikit-learn#24873)
56a31d6
jjerphan pushed a commit to jjerphan/scikit-learn that referenced this pull request Jan 23, 2023
@betatim @jjerphan
FIX Allow input validation by pass in MLPClassifier (scikit-learn#24873)
0ce78e0
adrinjalali pushed a commit that referenced this pull request Jan 24, 2023
@betatim @adrinjalali
FIX Allow input validation by pass in MLPClassifier (#24873)
c7eeb76
@betatim betatim mentioned this pull request Mar 17, 2023
Closed
@Charlie-XIAO Charlie-XIAO mentioned this pull request Mar 22, 2023
Merged
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

@glemaitre glemaitre glemaitre left review comments

Assignees
No one assigned
Labels
module:neural_network Validation related to input validation
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.

"X does not have valid feature names" when training on DataFrame input with MLP
4 participants
@betatim @cmarmo @glemaitre @thomasjpfan
0