skactiveml.pool.cost_reduction

skactiveml.pool.cost_reduction(k_vec_list, C=None, m_max=2, prior=0.001)

Calculate the expected cost reduction.

Calculate the expected cost reduction for given maximum number of hypothetically acquired labels, observed labels and cost matrix.

Parameters

k_vec_list: array-like, shape (n_samples, n_classes)

Observed class labels.

C: array-like, shape = (n_classes, n_classes)

Cost matrix.

m_max: int

Maximal number of hypothetically acquired labels.

priorfloat | array-like, shape (n_classes)

Prior value for each class.

Returns

expected_cost_reduction: array-like, shape (n_samples)

Expected cost reduction for given parameters.