MAINT: minor fixes to GaussianNB epsilon

scikit-learn · Oct 7, 2015 · eac920d · eac920d
1 parent e2e33aa
commit eac920d
Show file tree

Hide file tree

Showing 2 changed files with 3 additions and 3 deletions.
diff --git a/sklearn/naive_bayes.py b/sklearn/naive_bayes.py
@@ -324,7 +324,7 @@ def _partial_fit(self, X, y, classes=None, _refit=False,
         # will cause numerical errors. To address this, we artificially
         # boost the variance by epsilon, a small fraction of the standard
         # deviation of the largest dimension.
-        epsilon = 1e-9 * np.var(X, 0).max()
+        epsilon = 1e-9 * np.var(X, axis=0).max()
 
         if _refit:
             self.classes_ = None

diff --git a/sklearn/tests/test_naive_bayes.py b/sklearn/tests/test_naive_bayes.py
@@ -473,6 +473,6 @@ def test_naive_bayes_scale_invariance():
     iris = load_iris()
     X, y = iris.data, iris.target
     labels = [GaussianNB().fit(f * X, y).predict(f * X)
-              for f in [1E-5, 1, 1E5]]
+              for f in [1E-10, 1, 1E10]]
     assert_array_equal(labels[0], labels[1])
-    assert_array_equal(labels[0], labels[2])
+    assert_array_equal(labels[1], labels[2])