Module trustML.metrics.equalopportunity

Expand source code 
import numpy as np
from trustML.metrics.metric import Metric
from sklego.metrics import equal_opportunity_score

class EqualOpportunitySKL(Metric):
    """Equal opportunity metric of a sklearn-based classifier using sklego. The equality opportunity 
    score calculates the ratio between the probability of a **true positive** outcome given the sensitive 
    attribute (column) being true and the same probability given the sensitive attribute being false.

    This is especially useful to use in situations where "fairness" is a theme.

    (Extracted from sklego documentation)

    ADDITIONAL PROPERTIES: 
    - protected_attributes (list of str): list of sensible features
    - positive_class (optional): privileged class (if present, 1 otherwise)

    Args:
        Metric (Class): Metric abstract class
    """
    
    def __init__(self, additional_properties):
        super().__init__()
        self.protected_attributes = additional_properties["protected_attributes"]
        if "positive_class" in additional_properties:
            self.positive_class = additional_properties["positive_class"]
        else:
            self.positive_class = 1

    def assess(self, trained_model, data_x, data_y):    
        print("Computing equal opportunity metric...")
        if (self.protected_attributes is None):
            self.score = 1
        else:
            eq_opp_vector = np.zeros(len(self.protected_attributes))
            for i in range(len(self.protected_attributes)):
                eq_opp_vector[i] = equal_opportunity_score(
                    sensitive_column = self.protected_attributes[i], 
                    positive_target=self.positive_class)(trained_model, data_x, data_y)
                                  
            self.score = np.mean(eq_opp_vector)

Classes

class EqualOpportunitySKL (additional_properties)

Equal opportunity metric of a sklearn-based classifier using sklego. The equality opportunity score calculates the ratio between the probability of a true positive outcome given the sensitive attribute (column) being true and the same probability given the sensitive attribute being false.

This is especially useful to use in situations where "fairness" is a theme.

(Extracted from sklego documentation)

ADDITIONAL PROPERTIES: - protected_attributes (list of str): list of sensible features - positive_class (optional): privileged class (if present, 1 otherwise)

Args

Metric : Class
Metric abstract class
Expand source code 
class EqualOpportunitySKL(Metric):
    """Equal opportunity metric of a sklearn-based classifier using sklego. The equality opportunity 
    score calculates the ratio between the probability of a **true positive** outcome given the sensitive 
    attribute (column) being true and the same probability given the sensitive attribute being false.

    This is especially useful to use in situations where "fairness" is a theme.

    (Extracted from sklego documentation)

    ADDITIONAL PROPERTIES: 
    - protected_attributes (list of str): list of sensible features
    - positive_class (optional): privileged class (if present, 1 otherwise)

    Args:
        Metric (Class): Metric abstract class
    """
    
    def __init__(self, additional_properties):
        super().__init__()
        self.protected_attributes = additional_properties["protected_attributes"]
        if "positive_class" in additional_properties:
            self.positive_class = additional_properties["positive_class"]
        else:
            self.positive_class = 1

    def assess(self, trained_model, data_x, data_y):    
        print("Computing equal opportunity metric...")
        if (self.protected_attributes is None):
            self.score = 1
        else:
            eq_opp_vector = np.zeros(len(self.protected_attributes))
            for i in range(len(self.protected_attributes)):
                eq_opp_vector[i] = equal_opportunity_score(
                    sensitive_column = self.protected_attributes[i], 
                    positive_target=self.positive_class)(trained_model, data_x, data_y)
                                  
            self.score = np.mean(eq_opp_vector)

Ancestors

Inherited members