Metric tooltips and widget resizing changes

orangecontrib/fairness/evaluation/scoring.py

@@ -99,8 +99,8 @@ class StatisticalParityDifference(FairnessScorer):
 
  name = "SPD"
  long_name = str(
- "<p>Statistical Parity Difference (SPD): Measures the difference in ratios of "
- "favorable outcomes. An ideal value is 0.0.</p>"
+ "<p>Statistical Parity Difference (SPD): The difference in favorable "
+ "outcomes proportions between groups. An ideal value is 0.0.</p>"
  "<ul>"
  "<li>SPD &lt; 0: The privileged group has a higher rate of favorable outcomes.</li>"
  "<li>SPD &gt; 0: The privileged group has a lower rate of favorable outcomes.</li>"
@@ -118,9 +118,8 @@ class EqualOpportunityDifference(FairnessScorer):
 
  name = "EOD"
  long_name = str(
- "<p>Equal Opportunity Difference (EOD): It measures the difference in "
- "true positive rates. An ideal value is 0.0, indicating the difference "
- "in true positive rates is the same for both groups.</p>"
+ "<p>Equal Opportunity Difference (EOD): The difference in true positive rates between "
+ "groups. An ideal value is 0.0 meaning both groups have the same true positive rate.</p>"
  "<ul>"
  "<li>EOD &lt; 0: The privileged group has a higher true positive rate.</li>"
  "<li>EOD &gt; 0: The privileged group has a lower true positive rate.</li>"
@@ -138,11 +137,9 @@ class AverageOddsDifference(FairnessScorer):
 
  name = "AOD"
  long_name = str(
- "<p>Average Odds Difference (AOD): This metric calculates the average difference "
- "between the true positive rates (correctly predicting a positive outcome) and false "
- "positive rates (incorrectly predicting a positive outcome) for both the privileged "
- "and unprivileged groups. A value of 0.0 indicates equal rates for both groups, "
- "signifying fairness.</p>"
+ "<p>Average Odds Difference (AOD): The average of the differences in true "
+ "and false positive rates between privileged and unprivileged groups. "
+ "A value of 0.0 indicates equal rates for both groups.</p>"
  "<ul>"
  "<li>AOD &lt; 0: Indicates bias in favor of the privileged group.</li>"
  "<li>AOD &gt; 0: Indicates bias against the privileged group.</li>"
@@ -160,20 +157,17 @@ class DisparateImpact(FairnessScorer):
 
  name = "DI"
  long_name = str(
- "<p>Disparate Impact (DI): The ratio of ratios of favorable outcomes for an unprivileged "
- "group to that of the privileged group. An ideal value of 1.0 means the ratio is "
- "the same for both groups.</p>"
+ "<p>Disparate Impact (DI) is the ratio of favorable outcome "
+ "proportions between an unprivileged and privileged group. "
+ "Value of 1.0 indicates that the ratio is equal for both groups.</p>"
  "<ul>"
- "<li>DI &lt; 1.0: The privileged group receives favorable outcomes at a higher rate "
- "than the unprivileged group.</li>"
- "<li>DI &gt; 1.0: The privileged group receives favorable outcomes at a lower rate "
- "than the unprivileged group.</li>"
+ "<li>DI &lt; 1.0: The privileged group has a higher rate of favorable outcomes.</li>"
+ "<li>DI &gt; 1.0: The privileged group has a lower rate of favorable outcomes.</li>"
  "</ul>"
  )
 
  # TODO: When using randomize, models sometimes predict the same class for all instances
  # This can lead to division by zero in the Disparate Impact score
- # (and untrue results for the other scores)
- # What is the best way to handle this?
+ # and untrue results for the other scores.
  def metric(self, classification_metric):
  return classification_metric.disparate_impact()

orangecontrib/fairness/modeling/adversarial.py

@@ -154,7 +154,6 @@ def _fit_model(self, data):
 
  # Fit storage and fit functions were modified to use a Table/Storage object
  # This is because it's the easiest way to get the domain, and meta attributes
- # TODO: Should I use the X,Y,W format instead of the table format ? (Same for the model)
  def fit(self, data: Table) -> AdversarialDebiasingModel:
  (
  standard_dataset,

orangecontrib/fairness/modeling/postprocessing.py

@@ -104,7 +104,6 @@ def incompatibility_reason(self, domain):
 
  # Fit storage and fit functions were modified to use a Table/Storage object
  # This is because it's the easiest way to get the domain, and meta attributes
- # TODO: Should I use the X,Y,W format instead of the table format ? (Same for the model)
  def fit_storage(self, data):
  if isinstance(data, Table):
  self.fit(data)

orangecontrib/fairness/widgets/owadversarialdebiasing.py

@@ -72,6 +72,8 @@ class OWAdversarialDebiasing(ConcurrentWidgetMixin, OWBaseLearner):
  icon = "icons/adversarial_debiasing.svg"
  priority = 30
 
+ resizing_enabled = True
+
  class Inputs(OWBaseLearner.Inputs):
  """Inputs for the widgets, which are the same as for the super class (Data, Preprocessor)"""
 
@@ -112,7 +114,6 @@ class Information(OWBaseLearner.Information):
  )
 
  # We define the settings we want to use
- # TODO: Should i use context/domain settings?
  hidden_layers_neurons = Setting(100)
  number_of_epochs = Setting(50)
  batch_size = Setting(128)

orangecontrib/fairness/widgets/owdatasetbias.py

@@ -37,6 +37,7 @@ class OWDatasetBias(OWWidget):
 
  want_control_area = False
  resizing_enabled = False
+ resizing_enabled = True
 
  class Inputs:
  """Input for the widget - dataset."""
@@ -74,26 +75,25 @@ def set_data(self, data: Optional[Table]) -> None:
  disparate_impact = dataset_metric.disparate_impact()
  statistical_parity_difference = dataset_metric.statistical_parity_difference()
  self.disparate_impact_label.setText(
- f"Disparate Impact (ideal = 1): {round(disparate_impact, 3)}"
+ f"Disparate Impact (ideal = 1): {round(disparate_impact, 3):.3f}"
  )
  self.disparate_impact_label.setToolTip(
- "<p>Disparate Impact (DI): Measures the ratio of the ratios of favorable class "
- "values for an unprivileged group to that of the privileged group. An ideal value "
- "of 1.0 means the ratio of favorable class values is the same for both groups.</p>"
+ "<p>Disparate Impact (DI) is the ratio of favorable outcome "
+ "proportions between an unprivileged and privileged group. "
+ "Value of 1.0 indicates that the ratio is equal for both groups.</p>"
  "<ul>"
- "<li>DI &lt; 1.0: The privileged group has a higher percentage of favorable class values.</li>"
- "<li>DI &gt; 1.0: The privileged group has a lower percentage of favorable class values.</li>"
+ "<li>DI &lt; 1.0: The privileged group has a higher rate of favorable outcomes.</li>"
+ "<li>DI &gt; 1.0: The privileged group has a lower rate of favorable outcomes.</li>"
  "</ul>"
  )
  self.statistical_parity_difference_label.setText(
- f"Statistical Parity Difference (ideal = 0): {round(statistical_parity_difference, 3)}"
+ f"Statistical Parity Difference (ideal = 0): {round(statistical_parity_difference, 3):.3f}"
  )
  self.statistical_parity_difference_label.setToolTip(
- "<p>Statistical Parity Difference (SPD): Measures the difference in ratios of "
- "favorable class values between the unprivileged and the privileged groups. An "
- "ideal value for this metric is 0.</p>"
+ "<p>Statistical Parity Difference (SPD): The difference in favorable "
+ "outcomes proportions between groups. An ideal value is 0.0.</p>"
  "<ul>"
- "<li>SPD &lt; 0: The privileged group has a higher percentage of favorable class values.</li>"
- "<li>SPD &gt; 0: The privileged group has a lower percentage of favorable class values.</li>"
+ "<li>SPD &lt; 0: The privileged group has a higher rate of favorable outcomes.</li>"
+ "<li>SPD &gt; 0: The privileged group has a lower rate of favorable outcomes.</li>"
  "</ul>"
  )

orangecontrib/fairness/widgets/owreweighing.py

@@ -41,7 +41,6 @@ def __call__(self, data):
  data = self.model.transform(data)
  return data.instance_weights
 
- # TODO: Check if this is ok
  InheritEq = True
 
 

orangecontrib/fairness/widgets/utils.py

@@ -309,7 +309,7 @@ def table_to_standard_dataset(data) -> None:
  data = Impute()(data)
 
  xdf, ydf, mdf = data.to_pandas_dfs()
- # Merge xdf and ydf TODO: Check if I need to merge mdf
+ # Merge xdf and ydf (currently not merging mdf because it is not used)
  # This dataframe consists of all the data, the categorical variables values are
  # represented with the index of the value in domain[attribute].values
  df = ydf.merge(xdf, left_index=True, right_index=True)