Fixed computation of coefficient of variation in classification_metri…

…cs (Trusted-AI#288)

c_v = sqrt(2*GE(alpha=2))

Signed-off-by: Eric Augustine <eaugustine011@gmail.com>
Co-authored-by: Samuel Hoffman <hoffman.sc@gmail.com>

aif360/aif360-r/tests/testthat/test-classification.R

@@ -29,10 +29,10 @@ test_that("running dataset test", {
   expect_equal(cm$average_abs_odds_difference(), 0.4250, tolerance=0.000603)
   expect_equal(cm$average_odds_difference(), -0.07545, tolerance=5.32e-05)
   expect_equal(class(cm$binary_confusion_matrix()), "list")
-  expect_equal(cm$between_all_groups_coefficient_of_variation(), 0.08035, tolerance=3.65e-05)
+  expect_equal(cm$between_all_groups_coefficient_of_variation(), 0.0568, tolerance=3.65e-05)
   expect_equal(cm$between_all_groups_generalized_entropy_index(alpha=2), 0.00161, tolerance=3.49e-06)
   expect_equal(cm$between_all_groups_theil_index(), 0.0016, tolerance=8.24e-06)
-  expect_equal(cm$coefficient_of_variation(), 0.8040, tolerance=0.00105)
+  expect_equal(cm$coefficient_of_variation(), 0.5685, tolerance=0.00105)
   expect_equal(cm$disparate_impact(), 1.629, tolerance=0.00063)
   expect_equal(cm$equal_opportunity_difference(), -0.5)
   expect_equal(cm$error_rate(), 0.45)

aif360/explainers/metric_json_explainer.py

@@ -73,7 +73,7 @@ def between_all_groups_coefficient_of_variation(self):
         response = OrderedDict((
             ("metric", "Between All Groups Coefficient Of Variation"),
             ("message", outcome),
-            ("description", "Computed as twice the square root of pairwise entropy between every pair of privileged and underprivileged groups."),
+            ("description", "Computed as the square root of twice the pairwise entropy between every pair of privileged and underprivileged groups with alpha = 2."),
             ("ideal", "The ideal value of this metric is 0.") #2.0"
         ))
         return json.dumps(response)
@@ -103,7 +103,7 @@ def between_group_coefficient_of_variation(self):
         response = OrderedDict((
             ("metric", "Between Group Coefficient Of Variation"),
             ("message", outcome),
-            ("description", "Computed as twice the square root of pairwise entropy between a given pair of privileged and underprivileged groups."),
+            ("description", "Computed as the square root of twice the pairwise entropy between a given pair of privileged and underprivileged groups with alpha = 2."),
             ("ideal", "The ideal value of this metric is 0.0") #2.0"
         ))
         return json.dumps(response)
@@ -123,7 +123,7 @@ def between_group_theil_index(self):
         response = OrderedDict((
             ("metric", "Between Group Theil Index"),
             ("message", outcome),
-            ("description", "Computed as the pairwise entropy between a given pair of privileged and underprivileged groups."),
+            ("description", "Computed as the pairwise entropy between a given pair of privileged and underprivileged groups with alpha = 1."),
             ("ideal", "The ideal value of this metric is 0.0") #1.0"
         ))
         return json.dumps(response)
@@ -133,7 +133,7 @@ def coefficient_of_variation(self):
         response = OrderedDict((
             ("metric", "Coefficient Of Variation"),
             ("message", outcome),
-            ("description", "Computed as twice the generalized entropy index."),
+            ("description", "Computed as the square root of twice the generalized entropy index with alpha = 2."),
             ("ideal", "The ideal value of this metric is 0.0") #2.0"
         ))
         return json.dumps(response)

aif360/metrics/classification_metric.py

@@ -777,10 +777,10 @@ def theil_index(self):
         return self.generalized_entropy_index(alpha=1)
 
     def coefficient_of_variation(self):
-        r"""The coefficient of variation is two times the square root of the
+        r"""The coefficient of variation is the square root of two times the
         :meth:`generalized_entropy_index` with :math:`\alpha = 2`.
         """
-        return 2 * np.sqrt(self.generalized_entropy_index(alpha=2))
+        return np.sqrt(2*self.generalized_entropy_index(alpha=2))
 
     def between_group_theil_index(self):
         r"""The between-group Theil index is the
@@ -789,11 +789,11 @@ def between_group_theil_index(self):
         return self.between_group_generalized_entropy_index(alpha=1)
 
     def between_group_coefficient_of_variation(self):
-        r"""The between-group coefficient of variation is two times the square
-        root of the :meth:`between_group_generalized_entropy_index` with
+        r"""The between-group coefficient of variation is the square
+        root of two times the :meth:`between_group_generalized_entropy_index` with
         :math:`\alpha = 2`.
         """
-        return 2*np.sqrt(self.between_group_generalized_entropy_index(alpha=2))
+        return np.sqrt(2*self.between_group_generalized_entropy_index(alpha=2))
 
     def between_all_groups_theil_index(self):
         r"""The between-group Theil index is the
@@ -803,11 +803,11 @@ def between_all_groups_theil_index(self):
         return self.between_all_groups_generalized_entropy_index(alpha=1)
 
     def between_all_groups_coefficient_of_variation(self):
-        r"""The between-group coefficient of variation is two times the square
-        root of the :meth:`between_all_groups_generalized_entropy_index` with
+        r"""The between-group coefficient of variation is the square
+        root of two times the :meth:`between_all_groups_generalized_entropy_index` with
         :math:`\alpha = 2`.
         """
-        return 2 * np.sqrt(self.between_all_groups_generalized_entropy_index(
+        return np.sqrt(2*self.between_all_groups_generalized_entropy_index(
             alpha=2))
 
     def differential_fairness_bias_amplification(self, concentration=1.0):

aif360/sklearn/metrics/metrics.py

@@ -592,7 +592,7 @@ def theil_index(b):
     return generalized_entropy_index(b, alpha=1)
 
 def coefficient_of_variation(b):
-    r"""The coefficient of variation is two times the square root of the
+    r"""The coefficient of variation is the square root of two times the
     :func:`generalized_entropy_index` with :math:`\alpha = 2`.
 
     Args:
@@ -601,7 +601,7 @@ def coefficient_of_variation(b):
     See also:
         :func:`generalized_entropy_index`
     """
-    return 2 * np.sqrt(generalized_entropy_index(b, alpha=2))
+    return np.sqrt(2 * generalized_entropy_index(b, alpha=2))
 
 
 # TODO: use sample_weight?