shapiq.game_theory

Game-theoretic concepts and algorithms.

This module provides exact computation of interaction indices via ExactComputer, Moebius transforms via MoebiusConverter, and utility functions for working with the various interaction index types defined in the shapiq framework.

Classes

ExactComputer	Computes a variety of game-theoretic values exactly.
MoebiusConverter	Computes a variety of game-theoretic concepts exactly from Möbius coefficients.

Functions

aggregate_base_interaction	Aggregates the basis interaction values into an efficient interaction index.
get_bernoulli_weights	Returns the bernoulli weights in the k-additive approximation via SII.
index_generalizes_sv	Checks if the given index generalizes the Shapley Value.
index_generalizes_bv	Checks if the given index generalizes the Banzhaf Value.
get_computation_index	Returns the base index of a given interaction index.
is_index_aggregated	Checks if the given index is an aggregated interaction index as denoted by a -.
is_empty_value_the_baseline	Check if empty prediction is the baseline.
egalitarian_least_core	Computes the egalitarian least-core for the underlying game represented through the parameters.