Raise an error at fitting time if the ratio do not make sense.

Author: Guillaume Lemaitre

unbalanced_dataset/base_sampler.py

@@ -115,6 +115,14 @@ def fit(self, X, y):
             print('{} classes detected: {}'.format(uniques.size,
                                                    self.stats_c_))
 
+        # Check if the ratio provided at initialisation make sense
+        if isinstance(self.ratio_, float):
+            if self.ratio_ < (self.stats_c_[self.min_c_] /
+                              self.stats_c_[self.maj_c_]):
+                raise RuntimeError('The ratio requested at initialisation'
+                                   ' should be greater or equal than the'
+                                   ' balancing ratio of the current data.')
+
         return self
 
     @abstractmethod

unbalanced_dataset/ensemble/tests/test_balance_cascade.py

@@ -73,6 +73,17 @@ def test_bc_fit_single_class():
     assert_raises(RuntimeError, bc.fit, X, y_single_class)
 
 
+def test_bc_fit_invalid_ratio():
+    """Test either if an error is raised when the balancing ratio to fit is
+    smaller than the one of the data"""
+
+    # Create the object
+    ratio = 1. / 10000.
+    bc = BalanceCascade(ratio=ratio, random_state=RND_SEED)
+    # Fit the data
+    assert_raises(RuntimeError, bc.fit, X, Y)
+
+
 def test_bc_fit():
     """Test the fitting method"""
 

unbalanced_dataset/ensemble/tests/test_easy_ensemble.py

@@ -73,6 +73,17 @@ def test_ee_fit_single_class():
     assert_raises(RuntimeError, ee.fit, X, y_single_class)
 
 
+def test_ee_fit_invalid_ratio():
+    """Test either if an error is raised when the balancing ratio to fit is
+    smaller than the one of the data"""
+
+    # Create the object
+    ratio = 1. / 10000.
+    ee = EasyEnsemble(ratio=ratio, random_state=RND_SEED)
+    # Fit the data
+    assert_raises(RuntimeError, ee.fit, X, Y)
+
+
 def test_ee_fit():
     """Test the fitting method"""
 

unbalanced_dataset/over_sampling/random_over_sampler.py

@@ -147,10 +147,11 @@ def transform(self, X, y):
 
             # Define the number of sample to create
             if self.ratio_ == 'auto':
-                num_samples = self.stats_c_[self.maj_c_] - self.stats_c_[key]
+                num_samples = int(self.stats_c_[self.maj_c_] -
+                                  self.stats_c_[key])
             else:
-                num_samples = ((self.ratio_ * self.stats_c_[self.maj_c_]) -
-                               self.stats_c_[key])
+                num_samples = int((self.ratio_ * self.stats_c_[self.maj_c_]) -
+                                  self.stats_c_[key])            
 
             # Pick some elements at random
             np.random.seed(self.rs_)

unbalanced_dataset/over_sampling/tests/test_random_over_sampler.py

@@ -68,6 +68,17 @@ def test_ros_fit_single_class():
     assert_raises(RuntimeError, ros.fit, X, y_single_class)
 
 
+def test_ros_fit_invalid_ratio():
+    """Test either if an error is raised when the balancing ratio to fit is
+    smaller than the one of the data"""
+
+    # Create the object
+    ratio = 1. / 10000.
+    ros = RandomOverSampler(ratio=ratio, random_state=RND_SEED)
+    # Fit the data
+    assert_raises(RuntimeError, ros.fit, X, Y)
+
+
 def test_ros_fit():
     """Test the fitting method"""
 

unbalanced_dataset/under_sampling/tests/test_cluster_centroids.py

@@ -71,6 +71,17 @@ def test_cc_fit_single_class():
     assert_raises(RuntimeError, cc.fit, X, y_single_class)
 
 
+def test_cc_fit_invalid_ratio():
+    """Test either if an error is raised when the balancing ratio to fit is
+    smaller than the one of the data"""
+
+    # Create the object
+    ratio = 1. / 10000.
+    cc = ClusterCentroids(ratio=ratio, random_state=RND_SEED)
+    # Fit the data
+    assert_raises(RuntimeError, cc.fit, X, Y)
+
+
 def test_cc_fit():
     """Test the fitting method"""
 

unbalanced_dataset/under_sampling/tests/test_instance_hardness_threshold.py

@@ -87,6 +87,18 @@ def test_iht_fit_single_class():
     assert_raises(RuntimeError, iht.fit, X, y_single_class)
 
 
+def test_iht_fit_invalid_ratio():
+    """Test either if an error is raised when the balancing ratio to fit is
+    smaller than the one of the data"""
+
+    # Create the object
+    ratio = 1. / 10000.
+    iht = InstanceHardnessThreshold(ESTIMATOR, ratio=ratio,
+                                    random_state=RND_SEED)
+    # Fit the data
+    assert_raises(RuntimeError, iht.fit, X, Y)
+
+
 def test_iht_fit():
     """Test the fitting method"""
 

unbalanced_dataset/under_sampling/tests/test_nearmiss_1.py

@@ -83,6 +83,17 @@ def test_nearmiss_fit_single_class():
     assert_raises(RuntimeError, nm1.fit, X, y_single_class)
 
 
+def test_nm_fit_invalid_ratio():
+    """Test either if an error is raised when the balancing ratio to fit is
+    smaller than the one of the data"""
+
+    # Create the object
+    ratio = 1. / 10000.
+    nm = NearMiss(ratio=ratio, random_state=RND_SEED)
+    # Fit the data
+    assert_raises(RuntimeError, nm.fit, X, Y)
+
+
 def test_nm1_fit():
     """Test the fitting method"""
 

unbalanced_dataset/under_sampling/tests/test_nearmiss_2.py

@@ -83,6 +83,17 @@ def test_nearmiss_fit_single_class():
     assert_raises(RuntimeError, nm2.fit, X, y_single_class)
 
 
+def test_nm_fit_invalid_ratio():
+    """Test either if an error is raised when the balancing ratio to fit is
+    smaller than the one of the data"""
+
+    # Create the object
+    ratio = 1. / 10000.
+    nm = NearMiss(ratio=ratio, random_state=RND_SEED)
+    # Fit the data
+    assert_raises(RuntimeError, nm.fit, X, Y)
+
+
 def test_nm2_fit():
     """Test the fitting method"""
 

unbalanced_dataset/under_sampling/tests/test_nearmiss_3.py

@@ -83,6 +83,17 @@ def test_nearmiss_fit_single_class():
     assert_raises(RuntimeError, nm3.fit, X, y_single_class)
 
 
+def test_nm_fit_invalid_ratio():
+    """Test either if an error is raised when the balancing ratio to fit is
+    smaller than the one of the data"""
+
+    # Create the object
+    ratio = 1. / 10000.
+    nm = NearMiss(ratio=ratio, random_state=RND_SEED)
+    # Fit the data
+    assert_raises(RuntimeError, nm.fit, X, Y)
+
+
 def test_nm3_fit():
     """Test the fitting method"""