EducationalTestingService / EducationalTestingService/skll
Add support for self-training meta-estimator
Open
enhancement
help wanted
sklearn-compatibility
- Dominant language
- Python
- Stars
- 561
- Forks
- 69
- PR merge metrics
- No merged PRs in 30d
Description
Nice way to take advantage of unlabeled data (based on Yarowski's old self-supervised WSD paper). See more [here](https://scikit-learn.org/stable/auto_examples/release_highlights/plot_release_highlights_0_24_0.html#new-self-training-meta-estimator).
Contributor guide
Assessment
This issue has not been assessed yet.