[MRG] FIX top_k_accuracy_score ignoring labels for "multiclass" case #19721
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
|
Would love to see this one merge, anything needed to make sure it crosses the finish line? |
thomasjpfan
reviewed
Apr 2, 2021
|
Do you need help resolving the merge conflict? I'm interested in making sure this lands before the next release :). |
Thank you for pointing that out, Done. I further improved a commit description and included a fixup commit in another commit to have a cleaner history. |
|
I guess this bugfix is not slated to land in the pending release? |
jnothman
approved these changes
Apr 24, 2021
There was a problem hiding this comment.
This LGTM. @glemaitre should it go in 0.24.2 as a fix to a new feature??
|
We can add it in the upcoming release. You only need to move the entry in |
Currently the last 2 parameters of the added test fail, because the labels are not considered to decide whether the target is "binary" or "multiclass". The labels parameters is only used in later steps.
If a problem is actually "multiclass", and not all classes are contained in the parameter `y_true` , the function fails, because the determined type is "binary". That decision makes sense, if the parameter labels is not passed. The problem is that the function also fails, if the parameter `labels` is passed, although it would be possible to determine the type of and the number of classes in conjunction with this parameter. This commit fixes that, by checking whether the `labels` parameter has been set and contains more than 2 classes, if the type has been determined to be "binary" in the previous step.
This is for the case where `labels` is an `ndarray`. Co-authored-by: Thomas J. Fan <thomasjpfan@gmail.com>
Co-authored-by: Thomas J. Fan <thomasjpfan@gmail.com>
Co-authored-by: Thomas J. Fan <thomasjpfan@gmail.com>
|
@glemaitre thanks for reviewing. The changelog has been moved to |
|
Thanks |
glemaitre
pushed a commit
to glemaitre/scikit-learn
that referenced
this pull request
Apr 26, 2021
glemaitre
pushed a commit
that referenced
this pull request
Apr 28, 2021
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
Same changes as #19300, which I accidentally closed, but can not reopen.
What does this implement/fix? Explain your changes.
See #19300 for these changes and review by @thomasjpfan. If wanted, I can copy that content here.
Any other comments?
I'm sorry for this extra work/confusion.