:orphan:




.. raw:: html

    <div class="sphx-glr-thumbnails">

.. thumbnail-parent-div-open

.. thumbnail-parent-div-close

.. raw:: html

    </div>


===========
Basic Usage
===========

Examples of how to fit, predict, and transform (when applicable) the data with each of the LVQ
algorithms.



.. raw:: html

    <div class="sphx-glr-thumbnails">

.. thumbnail-parent-div-open

.. raw:: html

    <div class="sphx-glr-thumbcontainer" tooltip="Example of how to fit the GLVQ `[1]`_ algorithm on the classic iris dataset.">

.. only:: html

  .. image:: /auto_examples/01_basic_usage/images/thumb/sphx_glr_plot_01_glvq_thumb.png
    :alt:

  :ref:`sphx_glr_auto_examples_01_basic_usage_plot_01_glvq.py`

.. raw:: html

      <div class="sphx-glr-thumbnail-title">Generalized LVQ (GLVQ)</div>
    </div>


.. raw:: html

    <div class="sphx-glr-thumbcontainer" tooltip="Example of how to use GMLVQ `[1]`_ on the classic iris dataset.">

.. only:: html

  .. image:: /auto_examples/01_basic_usage/images/thumb/sphx_glr_plot_02_gmlvq_thumb.png
    :alt:

  :ref:`sphx_glr_auto_examples_01_basic_usage_plot_02_gmlvq.py`

.. raw:: html

      <div class="sphx-glr-thumbnail-title">Generalized Matrix LVQ (GMLVQ)</div>
    </div>


.. raw:: html

    <div class="sphx-glr-thumbcontainer" tooltip="Example of how to use LGMLVQ `[1]`_ on the classic iris dataset.">

.. only:: html

  .. image:: /auto_examples/01_basic_usage/images/thumb/sphx_glr_plot_03_lgmlvq_thumb.png
    :alt:

  :ref:`sphx_glr_auto_examples_01_basic_usage_plot_03_lgmlvq.py`

.. raw:: html

      <div class="sphx-glr-thumbnail-title">Local Generalized Matrix LVQ (LGMLVQ)</div>
    </div>


.. raw:: html

    <div class="sphx-glr-thumbcontainer" tooltip="NanLVQ `[1]`_ refers to a extension that can be implemented for various distance functions. It uses the partial distance strategy to ignore any NaN values in the data. Another interpretation would be that it imputes the missing values with those of the prototypes. Hence, the distance will be zero, which results in a zero update for the feature containing the NaN value.">

.. only:: html

  .. image:: /auto_examples/01_basic_usage/images/thumb/sphx_glr_plot_04_NaNLVQ_thumb.png
    :alt:

  :ref:`sphx_glr_auto_examples_01_basic_usage_plot_04_NaNLVQ.py`

.. raw:: html

      <div class="sphx-glr-thumbnail-title">Not a Number LVQ (NaNLVQ)</div>
    </div>


.. thumbnail-parent-div-close

.. raw:: html

    </div>


==============
Pre-processing
==============

Any pre-processing, can be achieved by using the by sklearn's `pipelines`_.
Therefore, this section will not discuss the topic in detail but provide a basic example of how one
would do this using a model from the sklvq package.

.. _pipelines: https://scikit-learn.org/stable/modules/compose.html



.. raw:: html

    <div class="sphx-glr-thumbnails">

.. thumbnail-parent-div-open

.. raw:: html

    <div class="sphx-glr-thumbcontainer" tooltip="In these examples GMLVQ is used but the same applies to all the other algorithms. Also the `pipelines`_ feature is provided by scikit-learn and we therefore refer to scikit-learn&#x27;s documentation  for more details.">

.. only:: html

  .. image:: /auto_examples/02_pre_processing/images/thumb/sphx_glr_plot_01_pipeline_lvq_thumb.png
    :alt:

  :ref:`sphx_glr_auto_examples_02_pre_processing_plot_01_pipeline_lvq.py`

.. raw:: html

      <div class="sphx-glr-thumbnail-title">Pipelines</div>
    </div>


.. thumbnail-parent-div-close

.. raw:: html

    </div>


===============
Model Selection
===============

This section contains how one could use sklearn's `crossvalidation`_ and `gridsearch`_ methods in
combination with the models provided in the sklvq package.

.. _crossvalidation: https://scikit-learn.org/stable/modules/cross_validation.html
.. _gridsearch: https://scikit-learn.org/stable/modules/grid_search.html


.. raw:: html

    <div class="sphx-glr-thumbnails">

.. thumbnail-parent-div-open

.. raw:: html

    <div class="sphx-glr-thumbcontainer" tooltip="In all previous examples we showed the training performance of the models. However, in practice it is much more interesting how well a model performs on unseen data, i.e., the generalizability of the model. We can use `crossvalidation`_ for this.">

.. only:: html

  .. image:: /auto_examples/03_model_selection/images/thumb/sphx_glr_plot_01_crossvalidation_lvq_thumb.png
    :alt:

  :ref:`sphx_glr_auto_examples_03_model_selection_plot_01_crossvalidation_lvq.py`

.. raw:: html

      <div class="sphx-glr-thumbnail-title">Cross validation</div>
    </div>


.. raw:: html

    <div class="sphx-glr-thumbcontainer" tooltip="Cross validation  is not the whole story as it only can tell you  the expected performance of a single set of (hyper) parameters. Luckily  sklearn also provides a way of trying out multiple settings and return the CV scores for each of them. We can use `gridsearch`_ for this.">

.. only:: html

  .. image:: /auto_examples/03_model_selection/images/thumb/sphx_glr_plot_02_gridsearch_lvq_thumb.png
    :alt:

  :ref:`sphx_glr_auto_examples_03_model_selection_plot_02_gridsearch_lvq.py`

.. raw:: html

      <div class="sphx-glr-thumbnail-title">Grid Search</div>
    </div>


.. thumbnail-parent-div-close

.. raw:: html

    </div>


===================
Performance Metrics
===================



.. raw:: html

    <div class="sphx-glr-thumbnails">

.. thumbnail-parent-div-open

.. raw:: html

    <div class="sphx-glr-thumbcontainer" tooltip="In these examples GMLVQ is used but the same applies to all the other algorithms. However, not each solver provides the same variables. Additionally, the options &quot;lbfgs&quot; and &quot;bfgs&quot; are implemented in scipy and their callbacks are different from the others. See Scipy&#x27;s documentation for further information.">

.. only:: html

  .. image:: /auto_examples/04_performance_metrics/images/thumb/sphx_glr_plot_01_learning_rate_lvq_thumb.png
    :alt:

  :ref:`sphx_glr_auto_examples_04_performance_metrics_plot_01_learning_rate_lvq.py`

.. raw:: html

      <div class="sphx-glr-thumbnail-title">Learning Behaviour</div>
    </div>


.. raw:: html

    <div class="sphx-glr-thumbcontainer" tooltip="Example of plotting the ROC curve for a classification task.">

.. only:: html

  .. image:: /auto_examples/04_performance_metrics/images/thumb/sphx_glr_plot_02_roc_curve_thumb.png
    :alt:

  :ref:`sphx_glr_auto_examples_04_performance_metrics_plot_02_roc_curve.py`

.. raw:: html

      <div class="sphx-glr-thumbnail-title">Receiver Operating Characteristic Curve</div>
    </div>


.. thumbnail-parent-div-close

.. raw:: html

    </div>


=============
Customization
=============

The algorithms accept custom activation, discriminant and distance functions, as well as solvers.
Any customization with proper testing and documentation will be considered to be included in the
sklvq package. Please, create a pull request on github.

Custom models and objectives are also welcome. However, as they greatly impact all currently
implemented  parts and even might work completely differently there is no easy way to make this
similarly expandable as what currently is.


.. raw:: html

    <div class="sphx-glr-thumbnails">

.. thumbnail-parent-div-open

.. raw:: html

    <div class="sphx-glr-thumbcontainer" tooltip="Activation Functions">

.. only:: html

  .. image:: /auto_examples/05_customization/images/thumb/sphx_glr_plot_01_activations_thumb.png
    :alt:

  :ref:`sphx_glr_auto_examples_05_customization_plot_01_activations.py`

.. raw:: html

      <div class="sphx-glr-thumbnail-title">Activation Functions</div>
    </div>


.. raw:: html

    <div class="sphx-glr-thumbcontainer" tooltip="Distance Functions">

.. only:: html

  .. image:: /auto_examples/05_customization/images/thumb/sphx_glr_plot_02_distances_thumb.png
    :alt:

  :ref:`sphx_glr_auto_examples_05_customization_plot_02_distances.py`

.. raw:: html

      <div class="sphx-glr-thumbnail-title">Distance Functions</div>
    </div>


.. raw:: html

    <div class="sphx-glr-thumbcontainer" tooltip="Discriminant Functions">

.. only:: html

  .. image:: /auto_examples/05_customization/images/thumb/sphx_glr_plot_03_discriminants_thumb.png
    :alt:

  :ref:`sphx_glr_auto_examples_05_customization_plot_03_discriminants.py`

.. raw:: html

      <div class="sphx-glr-thumbnail-title">Discriminant Functions</div>
    </div>


.. raw:: html

    <div class="sphx-glr-thumbcontainer" tooltip="Solvers">

.. only:: html

  .. image:: /auto_examples/05_customization/images/thumb/sphx_glr_plot_04_solvers_thumb.png
    :alt:

  :ref:`sphx_glr_auto_examples_05_customization_plot_04_solvers.py`

.. raw:: html

      <div class="sphx-glr-thumbnail-title">Solvers</div>
    </div>


.. thumbnail-parent-div-close

.. raw:: html

    </div>


.. toctree::
   :hidden:
   :includehidden:


   /auto_examples/01_basic_usage/index.rst
   /auto_examples/02_pre_processing/index.rst
   /auto_examples/03_model_selection/index.rst
   /auto_examples/04_performance_metrics/index.rst
   /auto_examples/05_customization/index.rst


.. only:: html

  .. container:: sphx-glr-footer sphx-glr-footer-gallery

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

      :download:`Download all examples in Python source code: auto_examples_python.zip </auto_examples/auto_examples_python.zip>`

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

      :download:`Download all examples in Jupyter notebooks: auto_examples_jupyter.zip </auto_examples/auto_examples_jupyter.zip>`


.. only:: html

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

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