shapiq.game_theory.aggregate_base_interaction

shapiq.game_theory.aggregate_base_interaction(base_interactions, order=None)

Aggregates the basis interaction values into an efficient interaction index.

An example aggregation would be the transformation from SII values to k-SII values.

Parameters:

• base_interactions (InteractionValues) – The basis interaction values to aggregate.

• order (int | None) – The order of the aggregation. For example, the order of the k-SII aggregation. If None, the maximum order of the base interactions is used. Defaults to None.

Return type:

InteractionValues

Returns:

The aggregated interaction values.

Raises:

ValueError – If the order is smaller than 0.

Examples

>>> import numpy as np
>>> from shapiq.interaction_values import InteractionValues
>>> sii_values = InteractionValues(
...     n_players=3,
...     values=np.array([-0.1, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6]),
...     index="SII",
...     interaction_lookup={(): 0, (1,): 1, (2,): 2, (3,): 3, (1, 2): 4, (2, 3): 5, (1, 3): 6},
...     baseline_value=0,  # for SII, the baseline value must not be the same as the values of emptyset
...     min_order=0,
...     max_order=2,
... )
>>> k_sii_values = aggregate_base_interaction(sii_values)
>>> k_sii_values.index
'k-SII'
>>> k_sii_values.baseline_value
0
>>> k_sii_values.interaction_lookup
{(): 0, (1,): 1, (2,): 2, (3,): 3, (1, 2): 4, (2, 3): 5, (1, 3): 6}
>>> k_sii_values.max_order
2