.. DO NOT EDIT.
.. THIS FILE WAS AUTOMATICALLY GENERATED BY SPHINX-GALLERY.
.. TO MAKE CHANGES, EDIT THE SOURCE PYTHON FILE:
.. "auto_examples/basics/plot_parallel_computation.py"
.. LINE NUMBERS ARE GIVEN BELOW.

.. only:: html

    .. note::
        :class: sphx-glr-download-link-note

        :ref:`Go to the end <sphx_glr_download_auto_examples_basics_plot_parallel_computation.py>`
        to download the full example code.

.. rst-class:: sphx-glr-example-title

.. _sphx_glr_auto_examples_basics_plot_parallel_computation.py:


Parallel Computation with joblib
=================================

This example shows how to speed up ``shapiq`` explanations using parallel
computation via `joblib <https://joblib.readthedocs.io/>`_. We use a simple
synthetic game to keep runtime short.

.. GENERATED FROM PYTHON SOURCE LINES 9-16

.. code-block:: Python


    from __future__ import annotations

    import numpy as np

    import shapiq








.. GENERATED FROM PYTHON SOURCE LINES 17-20

Define a Synthetic Game
-----------------------
A lightweight callable that we can explain quickly.

.. GENERATED FROM PYTHON SOURCE LINES 20-35

.. code-block:: Python


    rng = np.random.default_rng(42)
    n_features = 8
    weights = rng.standard_normal(n_features)


    def synthetic_model(x: np.ndarray) -> np.ndarray:
        """Simple linear model with interaction term."""
        return x @ weights + 0.5 * x[:, 0] * x[:, 1]


    # Create synthetic data
    X_background = rng.standard_normal((100, n_features))
    X_test = rng.standard_normal((6, n_features))








.. GENERATED FROM PYTHON SOURCE LINES 36-38

Explain a Single Instance
--------------------------

.. GENERATED FROM PYTHON SOURCE LINES 38-53

.. code-block:: Python


    explainer = shapiq.Explainer(
        model=synthetic_model,
        data=X_background,
        index="k-SII",
        max_order=2,
        random_state=0,
    )
    print(f"Explainer type: {type(explainer).__name__}")

    iv = explainer.explain(X_test[0], budget=256)
    print(iv)

    shapiq.network_plot(interaction_values=iv, feature_names=[f"x{i}" for i in range(n_features)])




.. image-sg:: /auto_examples/basics/images/sphx_glr_plot_parallel_computation_001.png
   :alt: plot parallel computation
   :srcset: /auto_examples/basics/images/sphx_glr_plot_parallel_computation_001.png
   :class: sphx-glr-single-img


.. rst-class:: sphx-glr-script-out

 .. code-block:: none

    Explainer type: TabularExplainer
    InteractionValues(
        index=k-SII, max_order=2, min_order=0, estimated=False, estimation_budget=256,
        n_players=8, baseline_value=0.3123873157557045,
        Top 10 interactions:
            (5,): 0.5176617643254364
            (): 0.3123873157557045
            (0, 1): 0.03694416683658724
            (6,): -0.06933886084594183
            (0,): -0.11273767884210385
            (7,): -0.1365021776001595
            (1,): -0.2622014244845414
            (2,): -0.26702089174324456
            (4,): -0.4915673843261027
            (3,): -0.5965700344493814
    )

    (<Figure size 700x700 with 1 Axes>, <Axes: >)



.. GENERATED FROM PYTHON SOURCE LINES 54-57

Parallel Explanation of Multiple Instances
--------------------------------------------
Use ``n_jobs`` in :meth:`~shapiq.Explainer.explain_X` to parallelize.

.. GENERATED FROM PYTHON SOURCE LINES 57-61

.. code-block:: Python


    ivs = explainer.explain_X(X_test, budget=256, n_jobs=2)
    print(f"Computed {len(ivs)} explanations")





.. rst-class:: sphx-glr-script-out

 .. code-block:: none

    Computed 6 explanations




.. GENERATED FROM PYTHON SOURCE LINES 62-64

Global Feature Importance
--------------------------

.. GENERATED FROM PYTHON SOURCE LINES 64-66

.. code-block:: Python


    shapiq.plot.bar_plot(ivs, feature_names=[f"x{i}" for i in range(n_features)])



.. image-sg:: /auto_examples/basics/images/sphx_glr_plot_parallel_computation_002.png
   :alt: plot parallel computation
   :srcset: /auto_examples/basics/images/sphx_glr_plot_parallel_computation_002.png
   :class: sphx-glr-single-img


.. rst-class:: sphx-glr-script-out

 .. code-block:: none


    <Axes: xlabel='Attribution'>




.. rst-class:: sphx-glr-timing

   **Total running time of the script:** (0 minutes 2.522 seconds)


.. _sphx_glr_download_auto_examples_basics_plot_parallel_computation.py:

.. only:: html

  .. container:: sphx-glr-footer sphx-glr-footer-example

    .. container:: sphx-glr-download sphx-glr-download-jupyter

      :download:`Download Jupyter notebook: plot_parallel_computation.ipynb <plot_parallel_computation.ipynb>`

    .. container:: sphx-glr-download sphx-glr-download-python

      :download:`Download Python source code: plot_parallel_computation.py <plot_parallel_computation.py>`

    .. container:: sphx-glr-download sphx-glr-download-zip

      :download:`Download zipped: plot_parallel_computation.zip <plot_parallel_computation.zip>`