PermutationSamplingSII

Permutation-based SII/k-SII approximation using PermutationSamplingSII.

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

InteractionValues(
    index=k-SII, max_order=2, min_order=0, estimated=True, estimation_budget=177,
    n_players=8, baseline_value=0.0,
    Top 10 interactions:
        (0,): 0.6499999999999999
        (1,): 0.5500000000000002
        (2,): 0.2
        (3,): 0.10000000000000002
        (4,): 0.04999999999999994
        (4, 5): 2.220446049250313e-16
        (1, 4): -1.249000902703301e-16
        (5,): -0.10000000000000003
        (6,): -0.20000000000000007
        (7,): -0.3
)

Force plot

iv.plot_force(feature_names=feature_names)

Network plot

iv.plot_network(feature_names=feature_names)