Add VotingChainladder (resolves #115)

[cbalona]

No description provided.

[cbalona]

Tests seems to be failing due to an sklearn import error: ImportError: cannot import name 'delayed' from 'sklearn.utils.fixes' (/usr/share/miniconda/envs/cl_test/lib/python3.7/site-packages/sklearn/utils/fixes.py)

[jbogaardt]

This is an awesome PR! I wouldn't have been able to come up with something this good.

Seems to only need a resolution for:

1. the ImportError` - is it a specific version of sklearn that we need to limit to?

Not needed for the PR, but we should also:
2. Add the class to to docs/modules/classes.rst TOC so its docstring shows up in the API docs
3. Perhaps take your unit test and add a plot to get it as a stand-alone 'Example' in the /examples folder for those that browse the docs by the example gallery
4. Add ensemble.rst to docs/modules/index.rst so the docs show up.
4. Add it to README.rst on the main repo page

[cbalona]

This is an awesome PR! I wouldn't have been able to come up with something this good.

Thank you John! I am glad you like it.

Seems to only need a resolution for:

1. the ImportError` - is it a specific version of sklearn that we need to limit to?

I don't think we should limit as it seems that the need to import delayed from sklearn.utils.fixes is temporary, and sklearn intends to remove that function once this issue is resolved: joblib/joblib#1071

As a more permanent fix, I am using delayed from joblib instead--it should do the same thing and all tests pass on my side.

Not needed for the PR, but we should also:
2. Add the class to to docs/modules/classes.rst TOC so its docstring shows up in the API docs
3. Perhaps take your unit test and add a plot to get it as a stand-alone 'Example' in the /examples folder for those that browse the docs by the example gallery
4. Add ensemble.rst to docs/modules/index.rst so the docs show up.
4. Add it to README.rst on the main repo page

Agreed on all the above.

[codecov-io]

Codecov Report

Merging #116 (a55209a) into master (920502f) will increase coverage by 0.13%.
The diff coverage is 81.90%.

@@            Coverage Diff             @@
##           master     #116      +/-   ##
==========================================
+ Coverage   80.08%   80.22%   +0.13%     
==========================================
  Files          63       68       +5     
  Lines        3465     3570     +105     
  Branches      480      497      +17     
==========================================
+ Hits         2775     2864      +89     
- Misses        508      516       +8     
- Partials      182      190       +8     

Flag	Coverage Δ
unittests	80.22% <81.90%> (+0.13%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
chainladder/ensemble/base.py	77.77% <77.77%> (ø)
chainladder/ensemble/voting.py	78.87% <78.87%> (ø)
chainladder/__init__.py	87.50% <100.00%> (+0.83%)	⬆️
chainladder/core/base.py	89.26% <100.00%> (+0.10%)	⬆️
chainladder/ensemble/__init__.py	100.00% <100.00%> (ø)
chainladder/ensemble/tests/test_predict.py	100.00% <100.00%> (ø)
chainladder/ensemble/tests/test_voting.py	100.00% <100.00%> (ø)
chainladder/methods/base.py	91.46% <100.00%> (+2.57%)	⬆️
... and 5 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 920502f...a55209a. Read the comment docs.

[jbogaardt]

Fantastic contribution @cbalona. Thank you so much!

[RonRichman]

Thanks everyone - great contribution @cbalona!