[MRG] DOC mention the multi-class scheme used

imblearn/combine/smote_enn.py

@@ -132,7 +132,8 @@ class SMOTEENN(SamplerMixin):
     -----
     The method is presented in [1]_.
 
-    Supports mutli-class resampling.
+    Supports mutli-class resampling. Refer to SMOTE and ENN regarding the
+    scheme which used.
 
     Examples
     --------

imblearn/combine/smote_tomek.py

@@ -102,7 +102,8 @@ class SMOTETomek(SamplerMixin):
     -----
     The methos is presented in [1]_.
 
-    Supports mutli-class resampling.
+    Supports mutli-class resampling. Refer to SMOTE and TomekLinks regarding
+    the scheme which used.
 
     Examples
     --------

imblearn/ensemble/balance_cascade.py

@@ -90,7 +90,8 @@ class BalanceCascade(BaseEnsembleSampler):
     -----
     The method is described in [1]_.
 
-    Supports mutli-class resampling.
+    Supports mutli-class resampling. A one-vs.-rest scheme is used as
+    originally proposed in [1]_.
 
     Examples
     --------

imblearn/ensemble/easy_ensemble.py

@@ -59,7 +59,7 @@ class EasyEnsemble(BaseEnsembleSampler):
     -----
     The method is described in [1]_.
 
-    Supports mutli-class resampling.
+    Supports mutli-class resampling by sampling each class independently.
 
     Examples
     --------

imblearn/over_sampling/adasyn.py

@@ -63,10 +63,10 @@ class ADASYN(BaseOverSampler):
 
     Notes
     -----
-    Supports mutli-class resampling.
-
     The implementation is based on [1]_.
 
+    Supports mutli-class resampling. A one-vs.-rest scheme is used.
+
     Examples
     --------
 

imblearn/over_sampling/random_over_sampler.py

@@ -46,7 +46,7 @@ class RandomOverSampler(BaseOverSampler):
 
     Notes
     -----
-    Supports mutli-class resampling.
+    Supports mutli-class resampling by sampling each class independently.
 
     Examples
     --------

imblearn/over_sampling/smote.py

@@ -99,7 +99,8 @@ class SMOTE(BaseOverSampler):
     -----
     See the original papers: [1]_, [2]_, [3]_ for more details.
 
-    Supports mutli-class resampling.
+    Supports mutli-class resampling. A one-vs.-rest scheme is used as
+    originally proposed in [1]_.
 
     Examples
     --------

imblearn/under_sampling/prototype_generation/cluster_centroids.py

@@ -58,7 +58,7 @@ class ClusterCentroids(BaseUnderSampler):
 
     Notes
     -----
-    Supports mutli-class resampling.
+    Supports mutli-class resampling by sampling each class independently.
 
     Examples
     --------

imblearn/under_sampling/prototype_selection/condensed_nearest_neighbour.py

@@ -75,7 +75,8 @@ class CondensedNearestNeighbour(BaseCleaningSampler):
     -----
     The method is based on [1]_.
 
-    Supports mutli-class resampling.
+    Supports mutli-class resampling. A one-vs.-rest scheme is used when
+    sampling a class as proposed in [1]_.
 
     Examples
     --------

imblearn/under_sampling/prototype_selection/edited_nearest_neighbours.py

@@ -84,7 +84,8 @@ class EditedNearestNeighbours(BaseCleaningSampler):
     -----
     The method is based on [1]_.
 
-    Supports mutli-class resampling.
+    Supports mutli-class resampling. A one-vs.-rest scheme is used when
+    sampling a class as proposed in [1]_.
 
     Examples
     --------
@@ -272,7 +273,8 @@ class RepeatedEditedNearestNeighbours(BaseCleaningSampler):
 
     Notes
     -----
-    The method is based on [1]_.
+    The method is based on [1]_. A one-vs.-rest scheme is used when
+    sampling a class as proposed in [1]_.
 
     Supports mutli-class resampling.
 
@@ -482,7 +484,8 @@ class AllKNN(BaseCleaningSampler):
     -----
     The method is based on [1]_.
 
-    Supports mutli-class resampling.
+    Supports mutli-class resampling. A one-vs.-rest scheme is used when
+    sampling a class as proposed in [1]_.
 
     Examples
     --------

imblearn/under_sampling/prototype_selection/instance_hardness_threshold.py

@@ -98,7 +98,8 @@ class InstanceHardnessThreshold(BaseCleaningSampler):
     -----
     The method is based on [1]_.
 
-    Supports mutli-class resampling.
+    Supports mutli-class resampling. A one-vs.-rest scheme is used when
+    sampling a class as proposed in [1]_.
 
     Examples
     --------

imblearn/under_sampling/prototype_selection/neighbourhood_cleaning_rule.py

@@ -71,7 +71,10 @@ class NeighbourhoodCleaningRule(BaseCleaningSampler):
 
     Notes
     -----
-    Supports mutli-class resampling.
+    See the original paper: [1]_.
+
+    Supports mutli-class resampling. A one-vs.-rest scheme is used when
+    sampling a class as proposed in [1]_.
 
     Examples
     --------

imblearn/under_sampling/prototype_selection/one_sided_selection.py

@@ -74,7 +74,9 @@ class OneSidedSelection(BaseCleaningSampler):
     -----
     The method is based on [1]_.
 
-    Supports mutli-class resampling.
+    Supports mutli-class resampling. A one-vs.-one scheme is used when sampling
+    a class as proposed in [1]_. For each class to be sampled, all samples of
+    this class and the minority class are used during the sampling procedure.
 
     Examples
     --------

imblearn/under_sampling/prototype_selection/random_under_sampler.py

@@ -52,7 +52,7 @@ class RandomUnderSampler(BaseUnderSampler):
 
     Notes
     -----
-    Supports mutli-class resampling.
+    Supports mutli-class resampling by sampling each class independently.
 
     Examples
     --------

imblearn/under_sampling/prototype_selection/tomek_links.py

@@ -52,7 +52,8 @@ class TomekLinks(BaseCleaningSampler):
     -----
     This method is based on [1]_.
 
-    Supports mutli-class resampling.
+    Supports mutli-class resampling. A one-vs.-rest scheme is used as
+    originally proposed in [1]_.
 
     Examples
     --------