tsfresh 0.18.0

v0.18.0

on Python PyPI

Added Features

• Allow arbitrary rolling sizes (#766)
• Allow for multiclass significance tests (#762)
• Add multiclass option to RelevantFeatureAugmenter (#782)
• Addition of matrix_profile feature (#793)
• Added new query similarity counter feature (#798)
• Add root mean square feature (#813)

Bugfixes/Typos/Documentation:

• Do not send coverage of notebook tests to codecov (#759)
• Fix typos in notebook (#757, #780)
• Fix output format of make_forecasting_frame (#758)
• Fix badges and remove benchmark test
• Fix BY notebook plot (#760)
• Ts forecast example improvement (#763)
• Also surpress warnings in dask (#769)
• Update relevant_feature_augmenter.py (#779)
• Fix column names in quick_start.rst (#778)
• Improve relevance table function documentation (#781)
• Fixed #789 Typo in "how to add custom feature" (#790)
• Convert to the correct type on warnings (#799)
• Fix minor typos in the docs (#802)
• Add unwanted filetypes to gitignore (#819)
• Fix build and test failures (#815)
• Fix imputing docu (#800)
• Bump the scikit-learn version (#822)