fairness aware

FAIRNESS_AWARE

Source code in engines/contentFilterEngine/fairness_explainability/fairness_aware.py

class FAIRNESS_AWARE:
    def __init__(self):
        """
        Initialize the fairness-aware module.

        Attributes:
            fairness_metrics (dict): A dictionary to store calculated fairness metrics, 
            such as the distribution of recommendations across different user demographics.
        """
        self.fairness_metrics = {}

    def evaluate_fairness(self, recommendations: Dict[int, List[int]], user_attributes: pd.DataFrame) -> Dict[str, float]:
        """
        Evaluate the fairness of the recommendations across different user groups.

        Parameters:
            recommendations (dict): A dictionary mapping user IDs to lists of recommended item IDs.
            user_attributes (pd.DataFrame): A DataFrame containing user demographic information, 
            such as age, gender, and other relevant attributes.

        Returns:
            dict: A dictionary of fairness metrics, providing insights into how recommendations 
            are distributed across different user groups. For example, it may include the 
            distribution of recommendations by gender or age group.
        """
        # Example: Calculate the distribution of recommendations across gender
        gender_distribution = user_attributes['gender'].value_counts(normalize=True).to_dict()
        self.fairness_metrics['gender_distribution'] = gender_distribution
        return self.fairness_metrics

    def ensure_fairness(self, recommendations: Dict[int, List[int]], user_attributes: pd.DataFrame) -> Dict[int, List[int]]:
        """
        Adjust recommendations to ensure fairness across user groups.

        Parameters:
            recommendations (dict): A dictionary mapping user IDs to lists of recommended item IDs.
            user_attributes (pd.DataFrame): A DataFrame containing user demographic information, 
            such as age, gender, and other relevant attributes.

        Returns:
            dict: Adjusted recommendations ensuring fairness, potentially modifying the original 
            recommendations to achieve a more balanced distribution across user groups.
        """
        # Placeholder: Implement logic to adjust recommendations for fairness
        return recommendations

__init__()

Initialize the fairness-aware module.

Attributes:

Name	Type	Description
fairness_metrics	dict	A dictionary to store calculated fairness metrics,

Source code in engines/contentFilterEngine/fairness_explainability/fairness_aware.py

def __init__(self):
    """
    Initialize the fairness-aware module.

    Attributes:
        fairness_metrics (dict): A dictionary to store calculated fairness metrics, 
        such as the distribution of recommendations across different user demographics.
    """
    self.fairness_metrics = {}

ensure_fairness(recommendations, user_attributes)

Adjust recommendations to ensure fairness across user groups.

Parameters:

Name	Type	Description	Default
recommendations	dict	A dictionary mapping user IDs to lists of recommended item IDs.	required
user_attributes	DataFrame	A DataFrame containing user demographic information,	required

Returns:

Name	Type	Description
dict	Dict[int, List[int]]	Adjusted recommendations ensuring fairness, potentially modifying the original
	Dict[int, List[int]]	recommendations to achieve a more balanced distribution across user groups.

Source code in engines/contentFilterEngine/fairness_explainability/fairness_aware.py

def ensure_fairness(self, recommendations: Dict[int, List[int]], user_attributes: pd.DataFrame) -> Dict[int, List[int]]:
    """
    Adjust recommendations to ensure fairness across user groups.

    Parameters:
        recommendations (dict): A dictionary mapping user IDs to lists of recommended item IDs.
        user_attributes (pd.DataFrame): A DataFrame containing user demographic information, 
        such as age, gender, and other relevant attributes.

    Returns:
        dict: Adjusted recommendations ensuring fairness, potentially modifying the original 
        recommendations to achieve a more balanced distribution across user groups.
    """
    # Placeholder: Implement logic to adjust recommendations for fairness
    return recommendations

evaluate_fairness(recommendations, user_attributes)

Evaluate the fairness of the recommendations across different user groups.

Parameters:

Name	Type	Description	Default
recommendations	dict	A dictionary mapping user IDs to lists of recommended item IDs.	required
user_attributes	DataFrame	A DataFrame containing user demographic information,	required

Returns:

Name	Type	Description
dict	Dict[str, float]	A dictionary of fairness metrics, providing insights into how recommendations
	Dict[str, float]	are distributed across different user groups. For example, it may include the
	Dict[str, float]	distribution of recommendations by gender or age group.

Source code in engines/contentFilterEngine/fairness_explainability/fairness_aware.py

def evaluate_fairness(self, recommendations: Dict[int, List[int]], user_attributes: pd.DataFrame) -> Dict[str, float]:
    """
    Evaluate the fairness of the recommendations across different user groups.

    Parameters:
        recommendations (dict): A dictionary mapping user IDs to lists of recommended item IDs.
        user_attributes (pd.DataFrame): A DataFrame containing user demographic information, 
        such as age, gender, and other relevant attributes.

    Returns:
        dict: A dictionary of fairness metrics, providing insights into how recommendations 
        are distributed across different user groups. For example, it may include the 
        distribution of recommendations by gender or age group.
    """
    # Example: Calculate the distribution of recommendations across gender
    gender_distribution = user_attributes['gender'].value_counts(normalize=True).to_dict()
    self.fairness_metrics['gender_distribution'] = gender_distribution
    return self.fairness_metrics