fairlearn.reductions.Moment#

class fairlearn.reductions.Moment[source]#

Generic moment.

Our implementations of the reductions approach to fairness [1] make use of Moment objects to describe both the optimization objective and the fairness constraints imposed on the solution. This is an abstract class for all such objects.

Read more in the User Guide.

Added in version 0.3.0.

bound()[source]#

Return vector of fairness bound constraint the length of gamma.

Return type:

Series

default_objective()[source]#

Return the default objective for the moment.

Return type:

Moment

gamma(predictor)[source]#

Calculate the degree to which constraints are currently violated by the predictor.

Return type:

Series

load_data(X, y, *, sensitive_features=None)[source]#

Load a set of data for use by this object.

Parameters:

  • X (array) – The feature array

  • y (Series) – The label vector

  • sensitive_features (Series | None) – The sensitive feature vector (default None)

Return type:

None

project_lambda(lambda_vec)[source]#

Return the projected lambda values.

Return type:

Series

signed_weights(lambda_vec)[source]#

Return the signed weights.

Return type:

Series

property index: MultiIndex | Index#

Return a pandas (multi-)index listing the constraints.

property total_samples: int#

Return the number of samples in the data.