v0.18.0

• 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)