FEA add stratified k-fold iterators for splitting multilabel data #26423
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
MultilabelStratifiedKFold and RepeatedMultilabelStratifiedKFold for splitting multilabel data
MultilabelStratifiedKFold and RepeatedMultilabelStratifiedKFold for splitting multilabel data… minor bugs; update examples and docstrings correspondingly
… between largest and smallest test sizes; refactoring based on the fact that multilabel-indicator contains only positive/negative examples; refactored tests correspondingly
|
@glemaitre Would you have time for a review? I have refactored the code to avoid creating two new classes. Please see also #26423 (comment) |
|
Let's not rush here and move it for next release. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Reference Issues/PRs
Towards #25193.
What does this implement/fix? Explain your changes.
This PR adds support for
"multilabel-indicator"inStratifiedKFold, which implements the iterative stratification algorithm for multi-label classification (I partially referred to the implementation here). Correspondingly,RepeatedStratifiedKFoldwould now support"multilabel-indicator"as well. The paper that proposed iterative stratification is this and this video may be helpful for understanding the algorithm.StratifiedKFoldwith multi-label target is tested in the following aspects (except the basics):As for documentation:
modules/doc/cross_validation.rst.examples\model_selection\plot_cv_indices.py, though it is not implemented in the PR yet.Some other comments:
[["0", "1"], ["1", "0"]]is considered"multiclass-multioutput"instead of"multiclass-indicator". Is this the desired behavior?model_selection/tests/test_split.pymay need to be refactored... It seems too messy now.