Random doctest failure in statistical_inference tutorial, model_selection #3389

Closed
vene opened this Issue Jul 15, 2014 · 1 comment

Projects

None yet

2 participants

@vene
Member
vene commented Jul 15, 2014

This was triggered on Travis by a completely unrelated commit.
It seems that the tutorial gets a better-than-expected result on digits :)

The Travis configuration that failed is DISTRIB="conda" PYTHON_VERSION="2.6" INSTALL_MKL="false" NUMPY_VERSION="1.6.2" SCIPY_VERSION="0.11.0"

I: Seeding RNGs with 565480497
[...]
======================================================================
FAIL: Doctest: model_selection.rst
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/home/travis/anaconda/envs/testenv/lib/python2.6/doctest.py", line 2163, in runTest
    raise self.failureException(self.format_failure(new.getvalue()))
AssertionError: Failed doctest test for model_selection.rst
  File "/home/travis/build/scikit-learn/scikit-learn/doc/tutorial/statistical_inference/model_selection.rst", line 0

----------------------------------------------------------------------
File "/home/travis/build/scikit-learn/scikit-learn/doc/tutorial/statistical_inference/model_selection.rst", line 148, in model_selection.rst
Failed example:
    clf.best_score_                                  # doctest: +ELLIPSIS
Expected:
    0.924...
Got:
    0.93561368209255535
----------------------------------------------------------------------
File "/home/travis/build/scikit-learn/scikit-learn/doc/tutorial/statistical_inference/model_selection.rst", line 150, in model_selection.rst
Failed example:
    clf.best_estimator_.gamma == 1e-6
Expected:
    True
Got:
    False
@amueller
Member

Wait, are we grid-searching the gamma parameter of a linear SVM here? That doesn't explain the error in accuracy really but ... it seems somewhat odd...

@larsmans larsmans closed this in #4171 Jan 28, 2015
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment