We're currently refactoring sktime for a new release (see our v0.4.0 milestone for details).

Part of the refactoring is to implement consistent unit tests for all estimators (see #246). I've implemented a number of basic unit tests that ensure consistency for estimators in our framework and compatibility with scikit-learn.

Many of our estimators already pass the tests and some of them were easy to fix. However, others still need some reworking.

What I need you to do is:

• please review the changes in Package refactoring  #246 that affect your estimators/files and let me know if there's anything that I should not have changed
• please open a PR to implement the necessary fixes based on the refactor branch (or dev branch once it's merged)
• if you have reproducible benchmarking results for a given data set, it'd be great to add a unit test that ensures reproducibility

Let me know if you need help or are unable to make these changes.

You can find the list of estimators that need to be fixed below. The unit tests are implemented in sktime/tests/ and sktime/utils/testing/.