Merge remote-tracking branch 'origin/master' into is/428

Author: glemaitre

doc/whats_new/v0.0.4.rst

@@ -60,6 +60,9 @@ Bug fixes
   and thus to obtain a deterministic results when using the same random state.
   :issue:`447` by :user:`Guillaume Lemaitre <glemaitre>`.
 
+- Force to clone scikit-learn estimator passed as attributes to samplers.
+  :issue:`446` by :user:`Guillaume Lemaitre <glemaitre>`.
+
 Maintenance
 ...........
 

imblearn/combine/smote_enn.py

@@ -9,6 +9,7 @@
 import logging
 import warnings
 
+from sklearn.base import clone
 from sklearn.utils import check_X_y
 
 from ..base import SamplerMixin
@@ -103,7 +104,7 @@ def _validate_estimator(self):
         "Private function to validate SMOTE and ENN objects"
         if self.smote is not None:
             if isinstance(self.smote, SMOTE):
-                self.smote_ = self.smote
+                self.smote_ = clone(self.smote)
             else:
                 raise ValueError('smote needs to be a SMOTE object.'
                                  'Got {} instead.'.format(type(self.smote)))
@@ -116,7 +117,7 @@ def _validate_estimator(self):
 
         if self.enn is not None:
             if isinstance(self.enn, EditedNearestNeighbours):
-                self.enn_ = self.enn
+                self.enn_ = clone(self.enn)
             else:
                 raise ValueError('enn needs to be an EditedNearestNeighbours.'
                                  ' Got {} instead.'.format(type(self.enn)))

imblearn/combine/smote_tomek.py

@@ -10,6 +10,7 @@
 import logging
 import warnings
 
+from sklearn.base import clone
 from sklearn.utils import check_X_y
 
 from ..base import SamplerMixin
@@ -111,7 +112,7 @@ def _validate_estimator(self):
 
         if self.smote is not None:
             if isinstance(self.smote, SMOTE):
-                self.smote_ = self.smote
+                self.smote_ = clone(self.smote)
             else:
                 raise ValueError('smote needs to be a SMOTE object.'
                                  'Got {} instead.'.format(type(self.smote)))
@@ -124,7 +125,7 @@ def _validate_estimator(self):
 
         if self.tomek is not None:
             if isinstance(self.tomek, TomekLinks):
-                self.tomek_ = self.tomek
+                self.tomek_ = clone(self.tomek)
             else:
                 raise ValueError('tomek needs to be a TomekLinks object.'
                                  'Got {} instead.'.format(type(self.tomek)))

imblearn/ensemble/balance_cascade.py

@@ -8,7 +8,7 @@
 
 import numpy as np
 
-from sklearn.base import ClassifierMixin
+from sklearn.base import ClassifierMixin, clone
 from sklearn.neighbors import KNeighborsClassifier
 from sklearn.utils import check_random_state, safe_indexing
 from sklearn.model_selection import cross_val_predict
@@ -142,7 +142,7 @@ def _validate_estimator(self):
         if (self.estimator is not None and
                 isinstance(self.estimator, ClassifierMixin) and
                 hasattr(self.estimator, 'predict')):
-            self.estimator_ = self.estimator
+            self.estimator_ = clone(self.estimator)
         elif self.estimator is None:
             self.estimator_ = KNeighborsClassifier()
         else:

imblearn/over_sampling/smote.py

@@ -14,6 +14,7 @@
 
 from scipy import sparse
 
+from sklearn.base import clone
 from sklearn.svm import SVC
 from sklearn.utils import check_random_state, safe_indexing
 
@@ -448,7 +449,7 @@ def _validate_estimator(self):
         if self.svm_estimator is None:
             self.svm_estimator_ = SVC(random_state=self.random_state)
         elif isinstance(self.svm_estimator, SVC):
-            self.svm_estimator_ = self.svm_estimator
+            self.svm_estimator_ = clone(self.svm_estimator)
         else:
             raise_isinstance_error('svm_estimator', [SVC],
                                    self.svm_estimator)
@@ -698,7 +699,7 @@ def _validate_estimator(self):
                         self.svm_estimator == 'deprecated'):
                     self.svm_estimator_ = SVC(random_state=self.random_state)
                 elif isinstance(self.svm_estimator, SVC):
-                    self.svm_estimator_ = self.svm_estimator
+                    self.svm_estimator_ = clone(self.svm_estimator)
                 else:
                     raise_isinstance_error('svm_estimator', [SVC],
                                            self.svm_estimator)

imblearn/under_sampling/prototype_generation/cluster_centroids.py

@@ -11,6 +11,7 @@
 import numpy as np
 from scipy import sparse
 
+from sklearn.base import clone
 from sklearn.cluster import KMeans
 from sklearn.neighbors import NearestNeighbors
 from sklearn.utils import safe_indexing
@@ -113,7 +114,7 @@ def _validate_estimator(self):
             self.estimator_ = KMeans(
                 random_state=self.random_state, n_jobs=self.n_jobs)
         elif isinstance(self.estimator, KMeans):
-            self.estimator_ = self.estimator
+            self.estimator_ = clone(self.estimator)
         else:
             raise ValueError('`estimator` has to be a KMeans clustering.'
                              ' Got {} instead.'.format(type(self.estimator)))

imblearn/under_sampling/prototype_selection/condensed_nearest_neighbour.py

@@ -13,6 +13,7 @@
 
 from scipy.sparse import issparse
 
+from sklearn.base import clone
 from sklearn.neighbors import KNeighborsClassifier
 from sklearn.utils import check_random_state, safe_indexing
 
@@ -121,7 +122,7 @@ def _validate_estimator(self):
             self.estimator_ = KNeighborsClassifier(
                 n_neighbors=self.n_neighbors, n_jobs=self.n_jobs)
         elif isinstance(self.n_neighbors, KNeighborsClassifier):
-            self.estimator_ = self.n_neighbors
+            self.estimator_ = clone(self.n_neighbors)
         else:
             raise ValueError('`n_neighbors` has to be a int or an object'
                              ' inhereited from KNeighborsClassifier.'

imblearn/under_sampling/prototype_selection/instance_hardness_threshold.py

@@ -12,7 +12,7 @@
 
 import numpy as np
 
-from sklearn.base import ClassifierMixin
+from sklearn.base import ClassifierMixin, clone
 from sklearn.ensemble import RandomForestClassifier
 from sklearn.model_selection import StratifiedKFold
 from sklearn.utils import safe_indexing
@@ -117,7 +117,7 @@ def _validate_estimator(self):
         if (self.estimator is not None and
                 isinstance(self.estimator, ClassifierMixin) and
                 hasattr(self.estimator, 'predict_proba')):
-            self.estimator_ = self.estimator
+            self.estimator_ = clone(self.estimator)
         elif self.estimator is None:
             self.estimator_ = RandomForestClassifier(
                 random_state=self.random_state, n_jobs=self.n_jobs)

imblearn/under_sampling/prototype_selection/one_sided_selection.py

@@ -9,6 +9,8 @@
 from collections import Counter
 
 import numpy as np
+
+from sklearn.base import clone
 from sklearn.neighbors import KNeighborsClassifier
 from sklearn.utils import check_random_state, safe_indexing
 
@@ -114,7 +116,7 @@ def _validate_estimator(self):
             self.estimator_ = KNeighborsClassifier(
                 n_neighbors=self.n_neighbors, n_jobs=self.n_jobs)
         elif isinstance(self.n_neighbors, KNeighborsClassifier):
-            self.estimator_ = self.n_neighbors
+            self.estimator_ = clone(self.n_neighbors)
         else:
             raise ValueError('`n_neighbors` has to be a int or an object'
                              ' inhereited from KNeighborsClassifier.'

imblearn/utils/tests/test_validation.py

@@ -36,7 +36,8 @@ def test_check_neighbors_object():
     assert issubclass(type(estimator), KNeighborsMixin)
     assert estimator.n_neighbors == 2
     estimator = NearestNeighbors(n_neighbors)
-    assert estimator is check_neighbors_object(name, estimator)
+    estimator_cloned = check_neighbors_object(name, estimator)
+    assert estimator.n_neighbors == estimator_cloned.n_neighbors
     n_neighbors = 'rnd'
     with pytest.raises(ValueError, match="has to be one of"):
         check_neighbors_object(name, n_neighbors)