FIX nan bug in BaseLabelPropagation (#19271)

ThuWangzw · thomasjpfan · web-flow · commit cd06652c12ab · 2021-02-01T21:53:36.000+01:00
Co-authored-by: Thomas J. Fan &lt;thomasjpfan@gmail.com&gt;
diff --git a/doc/whats_new/v1.0.rst b/doc/whats_new/v1.0.rst
@@ -150,6 +150,13 @@ Changelog
   for non-English characters. :pr:`18959` by :user:`Zero <Zeroto521>`
   and :user:`wstates <wstates>`.
 
+:mod:`sklearn.semi_supervised`
+.................................
+
+- |Fix| Avoid NaN during label propagation in
+  :class:`~sklearn.semi_supervised.LabelPropagation`.
+  :pr:`19271` by :user:`Zhaowei Wang <ThuWangzw>`.
+
 Code and Documentation Contributors
 -----------------------------------
 
diff --git a/sklearn/semi_supervised/_label_propagation.py b/sklearn/semi_supervised/_label_propagation.py
@@ -279,6 +279,7 @@ def fit(self, X, y):
             if self._variant == 'propagation':
                 normalizer = np.sum(
                     self.label_distributions_, axis=1)[:, np.newaxis]
+                normalizer[normalizer == 0] = 1
                 self.label_distributions_ /= normalizer
                 self.label_distributions_ = np.where(unlabeled,
                                                      self.label_distributions_,
diff --git a/sklearn/semi_supervised/tests/test_label_propagation.py b/sklearn/semi_supervised/tests/test_label_propagation.py
@@ -157,15 +157,19 @@ def test_convergence_warning():
     assert_no_warnings(mdl.fit, X, y)
 
 
-def test_label_propagation_non_zero_normalizer():
+@pytest.mark.parametrize("LabelPropagationCls",
+                         [label_propagation.LabelSpreading,
+                          label_propagation.LabelPropagation])
+def test_label_propagation_non_zero_normalizer(LabelPropagationCls):
     # check that we don't divide by zero in case of null normalizer
     # non-regression test for
     # https://github.com/scikit-learn/scikit-learn/pull/15946
+    # https://github.com/scikit-learn/scikit-learn/issues/9292
     X = np.array([[100., 100.], [100., 100.], [0., 0.], [0., 0.]])
     y = np.array([0, 1, -1, -1])
-    mdl = label_propagation.LabelSpreading(kernel='knn',
-                                           max_iter=100,
-                                           n_neighbors=1)
+    mdl = LabelPropagationCls(kernel='knn',
+                              max_iter=100,
+                              n_neighbors=1)
     assert_no_warnings(mdl.fit, X, y)
 
 
