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

.. only:: html

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

        :ref:`Go to the end <sphx_glr_download_glr_examples_applications_plot_kmf.py>`
        to download the full example code or to run this example in your browser via Binder

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

.. _sphx_glr_glr_examples_applications_plot_kmf.py:


=======================
K-Means Featurization
=======================
Featurize data for boosting the prediction and 
force model to generalization 

.. GENERATED FROM PYTHON SOURCE LINES 9-11

.. code-block:: default

    # License: BSD-3-clause
    # Author: L.Kouadio 







.. GENERATED FROM PYTHON SOURCE LINES 12-18

KMeans Featurisation ( KMF) is a surrogate booster to predict permeability 
coefficient (k) before any drilling construction. Indeed, KMF creates a 
compressed spatial index of the data which can be fed into the model for ease 
of learning and enforce the model capability of generalization. A new 
predictor based on model stacking technique is built with full target k 
which balances the spatial distribution of k-labels by clustering the original data. 

.. GENERATED FROM PYTHON SOURCE LINES 20-21

We start by importing the required modules 

.. GENERATED FROM PYTHON SOURCE LINES 21-29

.. code-block:: default

    import matplotlib.pyplot as plt

    from sklearn.datasets import make_moons
    from watex.transformers import featurize_X,  KMeansFeaturizer
    from watex.exlib import train_test_split 
    from watex.utils import plot_voronoi 









.. GENERATED FROM PYTHON SOURCE LINES 30-33

* Build models with a common data sets. (e.g. Moons dataset) 
We generate 8000 samples of dataset where we divided as 50% training and 50% testing 
For reproducing the same samples of data, we fixed the `seed`. 

.. GENERATED FROM PYTHON SOURCE LINES 33-38

.. code-block:: default

    seed = 1 # to reproduce the dataset
    X0 , y0 = make_moons(n_samples = 8000, noise= 0.2 , random_state= seed ) 
    X0_train, X0_test, y0_train, y0_test = train_test_split (
        X0, y0 , test_size =.5, random_state = seed )








.. GENERATED FROM PYTHON SOURCE LINES 39-43

Here, there is a shortcut way to featurize data at once by calling  the 
transformer :func:`watex.transformers.featurize_X` to transform X data. It 
could also returns  KMF_model (kmf_hint) if the parameter `return_model` is set 
to ``True``.  

.. GENERATED FROM PYTHON SOURCE LINES 43-46

.. code-block:: default

    X0_train_kmf, y0_train_kmf  = featurize_X(
        X0_train, y0_train , n_clusters =200, target_scale =10) 








.. GENERATED FROM PYTHON SOURCE LINES 47-51

* Voronoi plot 2D 
Veronoi plot can be used to visualize the model using hint ( target associated) and without. 
For a human representation ( 2D), we used the most two features importances 
of the consistent data set.

.. GENERATED FROM PYTHON SOURCE LINES 51-62

.. code-block:: default

 
    # Xpca = nPCA( X0_train, n_components= 2 ) # reduce the data set for two most feature components 
    # X0_test, y0_test = make_moons(n_samples=2000, noise=0.3)
    fig, ax = plt.subplots(2,1, figsize =(7, 7)) 
    kmf_hint = KMeansFeaturizer(n_clusters=200, target_scale=10).fit(X0_train,y0_train)
    kmf_no_hint = KMeansFeaturizer(n_clusters=200, target_scale=0).fit(X0_train, y0_train)
    plot_voronoi ( X0_train, y0_train ,cluster_centers=kmf_hint.cluster_centers_, 
                      fig_title ='KMF with hint', ax = ax [0] )
    plot_voronoi ( X0_train, y0_train,cluster_centers=kmf_no_hint.cluster_centers_, 
                      fig_title ='KMF No hint' , ax = ax[1])




.. image-sg:: /glr_examples/applications/images/sphx_glr_plot_kmf_001.png
   :alt: KMF with hint, KMF No hint
   :srcset: /glr_examples/applications/images/sphx_glr_plot_kmf_001.png
   :class: sphx-glr-single-img


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

 .. code-block:: none


    <AxesSubplot:title={'center':'KMF No hint'}>



.. GENERATED FROM PYTHON SOURCE LINES 63-66

As shown in the figure above. The number of clusters when target information is 
missed span too much of the space between the two classes. Commonly KMF demonstrates 
its usefulness when cluster boundaries align with class boundaries much more closely. 


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

   **Total running time of the script:** ( 0 minutes  8.712 seconds)


.. _sphx_glr_download_glr_examples_applications_plot_kmf.py:

.. only:: html

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


    .. container:: binder-badge

      .. image:: images/binder_badge_logo.svg
        :target: https://mybinder.org/v2/gh/watex/watex/master?urlpath=lab/tree/notebooks/glr_examples/applications/plot_kmf.ipynb
        :alt: Launch binder
        :width: 150 px



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

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

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

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


.. only:: html

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

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