Fix/label distribution entropy#733
Open
michaelellis003 wants to merge 2 commits intohuggingface:mainfrom
Open
Conversation
Skewness is statistically inappropriate for categorical label variables because it depends on arbitrary integer encoding and measures symmetry rather than uniformity. Replace it with Shannon entropy, which is permutation-invariant and correctly quantifies how balanced a label distribution is. Changes: - Replace scipy.stats.skew with scipy.stats.entropy - Return label_entropy (nats) and label_entropy_normalized (0 to 1) - Remove unused pandas import and string-to-integer conversion - Update docstrings, README examples, and references - Add test suite covering uniformity, imbalance, permutation invariance, string labels, and edge cases Fixes huggingface#659
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Fixes #659
Skewness is statistically inappropriate for categorical label variables — it depends on arbitrary integer encoding and measures symmetry rather than uniformity. For example, [0,0,1,1,1,1,1,2,2] and [0,0,1,1,2,2,2,2,2] have the same class distribution (2, 5, 2) but different skewness values. Entropy is identical for both, as expected.
This PR replaces label_skew with:
Note: this is a breaking change — label_skew is removed, not deprecated. The old value was statistically meaningless, so keeping it as a deprecated field would mean continuing to return a wrong number.
Changes: