kADDSHAP

k-additive SHAP approximation using kADDSHAP Pelegrina et al.[1].

from __future__ import annotations

import numpy as np

import shapiq

N_PLAYERS = 8
BUDGET = 200
feature_names = [f"x{i}" for i in range(N_PLAYERS)]

weights = np.array([0.4, 0.3, 0.2, 0.1, 0.05, -0.1, -0.2, -0.3])


def game_fun(coalitions: np.ndarray) -> np.ndarray:
    coalitions = np.atleast_2d(coalitions)
    return (coalitions @ weights) + 0.5 * coalitions[:, 0] * coalitions[:, 1]

Approximate k-SII values

approximator = shapiq.kADDSHAP(n=N_PLAYERS, max_order=2, random_state=42)
iv = approximator.approximate(BUDGET, game_fun)
print(iv)

InteractionValues(
    index=kADD-SHAP, max_order=2, min_order=0, estimated=True, estimation_budget=200,
    n_players=8, baseline_value=0.0,
    Top 10 interactions:
        (0,): 0.6499998956016835
        (1,): 0.5499999346814884
        (0, 1): 0.5000000131359702
        (2,): 0.19999997774749512
        (3,): 0.09999997907336332
        (4,): 0.05000002063526698
        (1, 6): 5.373277237993505e-08
        (5,): -0.09999993643380127
        (6,): -0.19999995674693913
        (7,): -0.29999991455855657
)

Force plot

iv.plot_force(feature_names=feature_names)

Network plot

iv.plot_network(feature_names=feature_names)

References

[1]

Guilherme Dean Pelegrina, Leonardo Tomazeli Duarte, and Michel Grabisch. A \emph k-additive choquet integral-based approach to approximate the SHAP values for local interpretability in machine learning. Artif. Intell., 325:104014, 2023. doi:10.1016/J.ARTINT.2023.104014.