.. DO NOT EDIT.
.. THIS FILE WAS AUTOMATICALLY GENERATED BY SPHINX-GALLERY.
.. TO MAKE CHANGES, EDIT THE SOURCE PYTHON FILE:
.. "auto_examples/plot_multitask_lasso_cv.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_plot_multitask_lasso_cv.py>`
        to download the full example code.

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

.. _sphx_glr_auto_examples_plot_multitask_lasso_cv.py:


==============================================================
Run MultitaskLassoCV and compare performance with scikit-learn
==============================================================

The example runs the MultitaskLassoCV scikit-learn like estimator.

.. GENERATED FROM PYTHON SOURCE LINES 8-19

.. code-block:: Python

    import time
    import numpy as np
    import matplotlib.pyplot as plt

    from celer import MultiTaskLassoCV
    from numpy.linalg import norm
    from sklearn.utils import check_random_state
    from sklearn import linear_model

    rng = check_random_state(0)








.. GENERATED FROM PYTHON SOURCE LINES 20-21

Generate some 2D coefficients with sine waves with random frequency and phase

.. GENERATED FROM PYTHON SOURCE LINES 21-35

.. code-block:: Python

    n_samples, n_features, n_tasks = 100, 500, 50
    n_relevant_features = 50
    support = rng.choice(n_features, n_relevant_features, replace=False)
    coef = np.zeros((n_tasks, n_features))
    times = np.linspace(0, 2 * np.pi, n_tasks)
    for k in support:
        coef[:, k] = np.sin((1. + rng.randn(1)) * times + 3 * rng.randn(1))


    X = rng.randn(n_samples, n_features)
    Y = np.dot(X, coef.T) + rng.randn(n_samples, n_tasks)
    Y /= norm(Y, ord='fro')









.. GENERATED FROM PYTHON SOURCE LINES 36-37

Fit with sklearn and celer, using the same API

.. GENERATED FROM PYTHON SOURCE LINES 37-46

.. code-block:: Python

    params = dict(tol=1e-6, cv=4, n_jobs=1, n_alphas=20)
    t0 = time.perf_counter()
    clf = MultiTaskLassoCV(**params).fit(X, Y)
    t_celer = time.perf_counter() - t0

    t0 = time.perf_counter()
    clf_sklearn = linear_model.MultiTaskLassoCV(**params).fit(X, Y)
    t_sklearn = time.perf_counter() - t0








.. GENERATED FROM PYTHON SOURCE LINES 47-48

Celer is faster

.. GENERATED FROM PYTHON SOURCE LINES 48-51

.. code-block:: Python

    print("Time for celer  : %.2f s" % t_celer)
    print("Time for sklearn: %.2f s" % t_sklearn)





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

 .. code-block:: none

    Time for celer  : 20.63 s
    Time for sklearn: 23.65 s




.. GENERATED FROM PYTHON SOURCE LINES 52-53

Both packages find the same solution

.. GENERATED FROM PYTHON SOURCE LINES 53-59

.. code-block:: Python

    print("Celer's optimal regularizer  : %s" % clf.alpha_)
    print("Sklearn's optimal regularizer: %s" % clf_sklearn.alpha_)

    print("Relative norm difference between optimal coefs: %.2f %%" %
          (100 * norm(clf.coef_ - clf_sklearn.coef_) / norm(clf.coef_)))





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

 .. code-block:: none

    Celer's optimal regularizer  : 0.00134781376541042
    Sklearn's optimal regularizer: 0.00134781376541042
    Relative norm difference between optimal coefs: 0.00 %




.. GENERATED FROM PYTHON SOURCE LINES 60-72

.. code-block:: Python

    fig, axarr = plt.subplots(2, 1, constrained_layout=True)
    axarr[0].spy(clf.coef_, aspect="auto")
    axarr[0].xaxis.tick_bottom()
    axarr[0].set_title("celer")
    axarr[0].set_ylabel("tasks")
    axarr[1].spy(clf_sklearn.coef_, aspect="auto")
    axarr[1].xaxis.tick_bottom()
    axarr[1].set_title("sklearn")
    plt.suptitle("Sparsity patterns")
    plt.ylabel("tasks")
    plt.xlabel("features")
    plt.show(block=False)



.. image-sg:: /auto_examples/images/sphx_glr_plot_multitask_lasso_cv_001.png
   :alt: Sparsity patterns, celer, sklearn
   :srcset: /auto_examples/images/sphx_glr_plot_multitask_lasso_cv_001.png
   :class: sphx-glr-single-img






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

   **Total running time of the script:** (0 minutes 44.641 seconds)


.. _sphx_glr_download_auto_examples_plot_multitask_lasso_cv.py:

.. only:: html

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

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

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

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

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

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

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


.. only:: html

 .. rst-class:: sphx-glr-signature

    `Gallery generated by Sphinx-Gallery <https://sphinx-gallery.github.io>`_