[MRG] EHN make_imbalance handle multi-class

[glemaitre]

Reference Issue

closes #283

What does this implement/fix? Explain your changes.

make_imbalance will support multi-class and will have a uniform API with the sampler.

Any other comments?

[glemaitre]

Appveyor is a bit slower but this should be done.

@massich @chkoar ready for reviews

[codecov]

Codecov Report

Merging #312 into master will decrease coverage by 0.01%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master     #312      +/-   ##
==========================================
- Coverage   98.33%   98.32%   -0.02%     
==========================================
  Files          68       68              
  Lines        3910     3879      -31     
==========================================
- Hits         3845     3814      -31     
  Misses         65       65

Impacted Files	Coverage Δ
...mpling/prototype_selection/random_under_sampler.py	100% <ø> (ø)	⬆️
imblearn/datasets/imbalance.py	100% <100%> (ø)	⬆️
imblearn/utils/tests/test_validation.py	100% <100%> (ø)	⬆️
imblearn/datasets/tests/test_imbalance.py	100% <100%> (ø)
imblearn/base.py	97.67% <100%> (ø)	⬆️
imblearn/utils/validation.py	100% <100%> (ø)	⬆️

---

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 f695fb7...71eb634. Read the comment docs.

[glemaitre]

I need this part now so please review and we will fix it later on.