- adds threshold and max_categories parameter to RobustOrdinalEncoder
- Add weight of evidence encoder
- use named functions instead of lambdas in DateTimeDefintions because of pickle
- update sklearn dependency version to 0.23 and mlio version to 0.5
- OrdinalEncoder can output np.nan instead of n for unseen values
- minor performance optimizations and refactoring
- adds a
get_classes
method toRobustLabelEncoder
- Merge pull request #18 from ipanepen/rle-bug
- test data reading when n_rows = 1 mod batch_size
- bug fix: makes fit_transform behavior consistent with fit and transform
- fix a minor bug in OneHotEncoder by by overloading the buggy method in ThresholdOneHotEncoder and fixing it
- dummy feature commit for RobustOrdinalEncoder & add badges to README
- libprotobuf==3.11.4 is not backwards compatible, specify tox version for testing
- Merge pull request #11 from ipanepen/master
- fix for MemoryError in ThresholdOneHotEncoder
- Adding RobustOrdinalEncoder
- Specify mlio version 0.2.7
- update to 1.0.0, fix buildspec
- update ci deployment credentials
- Merge pull request #4 from wiltonwu/master
- update documentation, remove CHANGELOG.md for 0.1.0 deployment, add date_time module
- Merge pull request #2 from ipanepen/ipanepen-add-random-seed
- adds np.random.seed(0) to test_preprocessing.py to ensure deterministic behavior
- Initial commit