[ENH] Refactor ColumnEnsemble to ChannelEnsemble

[TonyBagnall]

This refactors the classifier ColumEnsembleClassifier to becomre ChannelEnsembleClassifier based on discussion on #121

Other changes

1. Internal names changed from column to channel
2. Base class BaseChannelEnsembleClassifier made private
3. Internal input type changed to numpy3D (see [ENH] removing the use of nested_univ/nested np.DataFrames in time series machine learning #110)
4. Removes the ability to give channel names

this could now be significantly simplified and tidied up. I see no real need for a base class and a lot of the checks relate only to "nested_univ" Pandas datatype. I think all that can be part of the introduction of lists of numpy for unequal length.

Reference Issues/PRs

Fixes #121
partially addresses #110

Please use keywords (e.g., Fixes) to create link to the issues or pull requests
you resolved, so that they will automatically be closed when your pull request
is merged. See https://github.com/blog/1506-closing-issues-via-pull-requests
-->

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[patrickzib]

Minor comments

[patrickzib]

Fine with me.