Add short explanation of regression test do contributing docs. #8679 0.17.x - #8813

…at it always does that.

RandomTreeEmbeddings overwrites the fit_transform method of
_LearntSelectorMixin

Residues can be empty if the rank of X does not satisfy the
conditions described in scipy.linalg.lstsq documentation.
As residues are not that useful (i.e. we're not doing
stat testing), this property is deprecated and will be
removed in sklearn 0.19.

…sparse case.

…nge to 0.16 whatsnew.

…h will fail in future versions of numpy)

that have been generated with scikit-learn < 0.17

[MRG] backport CI fixes to 0.17.X branch

[ci skip]

scipy.misc.lena will be removed in scipy version 0.17.
This changes the example to use scipy.misc.face instead.

Flip sign according to `u` in both cases of `transpose`.

Backport 0.17.X -- FIX in randomized_svd flip sign

Using a different resource than lib.stat.cmu.edu which has not been fixed yet.

* fix floating point indexing in generic univariate

* weird work-around for spearman test

* add test for auto increasing