Revert "Fix MultiOutputClassifier checking for predict_proba method o…

…f base estimator (scikit-learn#12222)"

This reverts commit 7248191.

doc/whats_new/v0.21.rst

@@ -506,14 +506,6 @@ Support for Python 3.4 and below has been officially dropped.
   containing this same sample due to the scaling used in decision_function.
   :issue:`10440` by :user:`Jonathan Ohayon <Johayon>`.
 
-:mod:`sklearn.multioutput`
-........................
-
-- |Fix| Fixed a bug in :class:`multiout.MultiOutputClassifier` where the
-  `predict_proba` method incorrectly checked for `predict_proba` attribute in
-  the estimator object.
-  :issue:`12222` by :user:`Rebekah Kim <rebekahkim>`
-
 :mod:`sklearn.neighbors`
 ........................
 

sklearn/multioutput.py

@@ -145,7 +145,7 @@ def fit(self, X, y, sample_weight=None):
 
         if not hasattr(self.estimator, "fit"):
             raise ValueError("The base estimator should implement"
-                             " a fit method")
+                             "  a fit method")
 
         X, y = check_X_y(X, y,
                          multi_output=True,
@@ -186,8 +186,7 @@ def predict(self, X):
         """
         check_is_fitted(self, 'estimators_')
         if not hasattr(self.estimator, "predict"):
-            raise ValueError("The base estimator should implement"
-                             " a predict method")
+            raise ValueError("The base estimator should implement a predict method")
 
         X = check_array(X, accept_sparse=True)
 
@@ -328,9 +327,6 @@ def predict_proba(self, X):
         """Probability estimates.
         Returns prediction probabilities for each class of each output.
 
-        This method will raise a ``ValueError`` if any of the
-        estimators do not have ``predict_proba``.
-
         Parameters
         ----------
         X : array-like, shape (n_samples, n_features)
@@ -344,17 +340,16 @@ def predict_proba(self, X):
             classes corresponds to that in the attribute `classes_`.
         """
         check_is_fitted(self, 'estimators_')
-        if not all([hasattr(estimator, "predict_proba")
-                    for estimator in self.estimators_]):
-            raise ValueError("The base estimator should implement "
+        if not hasattr(self.estimator, "predict_proba"):
+            raise ValueError("The base estimator should implement"
                              "predict_proba method")
 
         results = [estimator.predict_proba(X) for estimator in
                    self.estimators_]
         return results
 
     def score(self, X, y):
-        """Returns the mean accuracy on the given test data and labels.
+        """"Returns the mean accuracy on the given test data and labels.
 
         Parameters
         ----------

sklearn/tests/test_multioutput.py

@@ -31,7 +31,6 @@
 from sklearn.svm import LinearSVC
 from sklearn.base import ClassifierMixin
 from sklearn.utils import shuffle
-from sklearn.model_selection import GridSearchCV
 
 
 def test_multi_target_regression():
@@ -177,34 +176,6 @@ def test_multi_output_classification_partial_fit_parallelism():
         assert est1 is not est2
 
 
-# check predict_proba passes
-def test_multi_output_predict_proba():
-    sgd_linear_clf = SGDClassifier(random_state=1, max_iter=5, tol=1e-3)
-    param = {'loss': ('hinge', 'log', 'modified_huber')}
-
-    # inner function for custom scoring
-    def custom_scorer(estimator, X, y):
-        if hasattr(estimator, "predict_proba"):
-            return 1.0
-        else:
-            return 0.0
-    grid_clf = GridSearchCV(sgd_linear_clf, param_grid=param,
-                            scoring=custom_scorer, cv=3, error_score=np.nan)
-    multi_target_linear = MultiOutputClassifier(grid_clf)
-    multi_target_linear.fit(X, y)
-
-    multi_target_linear.predict_proba(X)
-
-    # SGDClassifier defaults to loss='hinge' which is not a probabilistic
-    # loss function; therefore it does not expose a predict_proba method
-    sgd_linear_clf = SGDClassifier(random_state=1, max_iter=5, tol=1e-3)
-    multi_target_linear = MultiOutputClassifier(sgd_linear_clf)
-    multi_target_linear.fit(X, y)
-    err_msg = "The base estimator should implement predict_proba method"
-    with pytest.raises(ValueError, match=err_msg):
-        multi_target_linear.predict_proba(X)
-
-
 # 0.23. warning about tol not having its correct default value.
 @pytest.mark.filterwarnings('ignore:max_iter and tol parameters have been')
 def test_multi_output_classification_partial_fit():