[MRG] EHN: split and factorize SMOTE classes #440
Conversation
|
Hello @glemaitre! Thanks for updating the PR. Cheers ! There are no PEP8 issues in this Pull Request. 🍻 Comment last updated on July 27, 2018 at 20:17 Hours UTC |
Codecov Report
@@ Coverage Diff @@
## master #440 +/- ##
==========================================
+ Coverage 98.71% 98.75% +0.04%
==========================================
Files 70 70
Lines 4188 4270 +82
==========================================
+ Hits 4134 4217 +83
+ Misses 54 53 -1
Continue to review full report at Codecov.
|
glemaitre
changed the title
|
@chkoar @StephanHeijl Could you have a look to this PR. |
| assert_allclose(X_resampled, X_gt, rtol=R_TOL) | ||
| assert_array_equal(y_resampled, y_gt) | ||
|
|
||
|
|
||
| @pytest.mark.filterwarnings('ignore:"kind" is deprecated in 0.4 and will be') |
There was a problem hiding this comment.
It looks like all these filter warnings require some extra text.
There was a problem hiding this comment.
This is a regular expression. Only the beginning is useful (mainly the name of the parameter)
| [1.07844561, -0.19435291], [1.44015515, -1.30621303]]) | ||
| y_gt = np.array( | ||
| [0, 1, 0, 0, 0, 1, 1, 1, 1, 1, 1, 0, 0, 1, 1, 1, 1, 0, 1, 0, 0, 0, 0]) | ||
| X_gt = np.array([[0.11622591, -0.0317206], |
There was a problem hiding this comment.
Not extremely important, but it might be good to pick a single format for X* arrays here; 10x2 or 20x1.
|
Just looked this over, the code was moved around properly (as evidenced by the succeeding tests), so I only found small remarks. Aside from those it looks good to me, I'll be happy to adjust the KMeans code to conform to this format. |
Split the SMOTE class into several classes.
TODO:
SVMSMOTEandBorderlineSMOTERelated to #435