:orphan:


.. _l-notebooks:

=================
Notebooks Gallery
=================



:ref:`l-notebooks-coverage`


.. contents::
    :depth: 1
    :local:


.. raw:: html

   <div style='clear:both'></div>


Exploration
===========

Notebooks about experimentations.

.. contents::
    :local:


.. toctree::
    :maxdepth: 1

    notebooks/search_images_keras
    notebooks/search_images_torch

.. raw:: html

    <div class="sphx-pyq-thumb" tooltip="Images are usually very different if we compare them at pixel level but that's quite different if we look at them after they were processed by a deep learning model. We convert each image into a feature vector extracted from an intermediate layer of the network.">

.. only:: html

    .. figure:: /notebooks/search_images_keras.thumb.png

        :ref:`searchimageskerasrst`

.. raw:: html

    </div>


.. raw:: html

    <div class="sphx-pyq-thumb" tooltip="Images are usually very different if we compare them at pixel level but that's quite different if we look at them after they were processed by a deep learning model. We convert each image into a feature vector extracted from an intermediate layer of the network.">

.. only:: html

    .. figure:: /notebooks/search_images_torch.thumb.png

        :ref:`searchimagestorchrst`

.. raw:: html

    </div>

.. raw:: html

   <div style='clear:both'></div>


Extensions to scikit-learn
==========================

The following notebooks shows machine learning
object which follow :epkg:`scikit-learn` API
not implemented in :epkg:`scikit-learn`.


.. toctree::
    :maxdepth: 1

    notebooks/decision_tree_logreg
    notebooks/faster_polynomial_features
    notebooks/kmeans_l1
    notebooks/logistic_regression_clustering
    notebooks/piecewise_classification
    notebooks/piecewise_linear_regression
    notebooks/predictable_tsne
    notebooks/quantile_mlpregression
    notebooks/quantile_regression
    notebooks/regression_confidence_interval
    notebooks/traceable_ngrams_tfidf
    notebooks/sklearn_transformed_target
    notebooks/visualize_pipeline

.. raw:: html

    <div class="sphx-pyq-thumb" tooltip="The notebook demonstrates the model *DecisionTreeLogisticRegression* which replaces the decision based on one variable by a logistic regression.">

.. only:: html

    .. figure:: /notebooks/decision_tree_logreg.thumb.png

        :ref:`decisiontreelogregrst`

.. raw:: html

    </div>


.. raw:: html

    <div class="sphx-pyq-thumb" tooltip="">

.. only:: html

    .. figure:: /notebooks/faster_polynomial_features.thumb.png

        :ref:`fasterpolynomialfeaturesrst`

.. raw:: html

    </div>


.. raw:: html

    <div class="sphx-pyq-thumb" tooltip="This demonstrates how results change when using norm L1 for a k-means algorithm.">

.. only:: html

    .. figure:: /notebooks/kmeans_l1.thumb.png

        :ref:`kmeansl1rst`

.. raw:: html

    </div>


.. raw:: html

    <div class="sphx-pyq-thumb" tooltip="A logistic regression implements a convex partition of the features spaces. A clustering algorithm applied before the trainer modifies the feature space in way the partition is not necessarily convex in the initial features. Let's see how.">

.. only:: html

    .. figure:: /notebooks/logistic_regression_clustering.thumb.png

        :ref:`logisticregressionclusteringrst`

.. raw:: html

    </div>


.. raw:: html

    <div class="sphx-pyq-thumb" tooltip="Piecewise regression is easier to understand but the concept can be extended to classification. That's what this notebook explores.">

.. only:: html

    .. figure:: /notebooks/piecewise_classification.thumb.png

        :ref:`piecewiseclassificationrst`

.. raw:: html

    </div>


.. raw:: html

    <div class="sphx-pyq-thumb" tooltip="The notebook illustrates an implementation of a piecewise linear regression based on scikit-learn. The bucketization can be done with a DecisionTreeRegressor or a KBinsDiscretizer. A linear model is then fitted on each bucket.">

.. only:: html

    .. figure:: /notebooks/piecewise_linear_regression.thumb.png

        :ref:`piecewiselinearregressionrst`

.. raw:: html

    </div>


.. raw:: html

    <div class="sphx-pyq-thumb" tooltip="t-SNE is not a transformer which can produce outputs for other inputs than the one used to train the transform. The proposed solution is train a predictor afterwards to try to use the results on some other inputs the model never saw.">

.. only:: html

    .. figure:: /notebooks/predictable_tsne.thumb.png

        :ref:`predictabletsnerst`

.. raw:: html

    </div>


.. raw:: html

    <div class="sphx-pyq-thumb" tooltip="scikit-learn does not have a quantile regression for multi-layer perceptron. mlinsights implements a version of it based on the *scikit-learn* model. The implementation overwrites method ``_backprop``.">

.. only:: html

    .. figure:: /notebooks/quantile_mlpregression.thumb.png

        :ref:`quantilemlpregressionrst`

.. raw:: html

    </div>


.. raw:: html

    <div class="sphx-pyq-thumb" tooltip="scikit-learn does not have a quantile regression. mlinsights implements a version of it.">

.. only:: html

    .. figure:: /notebooks/quantile_regression.thumb.png

        :ref:`quantileregressionrst`

.. raw:: html

    </div>


.. raw:: html

    <div class="sphx-pyq-thumb" tooltip="The notebook computes confidence intervals with bootstrapping and quantile regression on a simple problem.">

.. only:: html

    .. figure:: /notebooks/regression_confidence_interval.thumb.png

        :ref:`regressionconfidenceintervalrst`

.. raw:: html

    </div>


.. raw:: html

    <div class="sphx-pyq-thumb" tooltip="The notebook looks into the way n-grams are stored in CountVectorizer and TfidfVectorizer and how the current storage (<= 0.21) is ambiguous in some cases.">

.. only:: html

    .. figure:: /notebooks/traceable_ngrams_tfidf.thumb.png

        :ref:`traceablengramstfidfrst`

.. raw:: html

    </div>


.. raw:: html

    <div class="sphx-pyq-thumb" tooltip="TransformedTargetRegressor proposes a way to modify the target before training. The notebook extends the concept to classifiers.">

.. only:: html

    .. figure:: /notebooks/sklearn_transformed_target.thumb.png

        :ref:`sklearntransformedtargetrst`

.. raw:: html

    </div>


.. raw:: html

    <div class="sphx-pyq-thumb" tooltip="Pipeline can be big with *scikit-learn*, let's dig into a visual way to look a them.">

.. only:: html

    .. figure:: /notebooks/visualize_pipeline.thumb.png

        :ref:`visualizepipelinerst`

.. raw:: html

    </div>

.. raw:: html

   <div style='clear:both'></div>


Extensions to scikit-learn involving Cython
===========================================

Experiments with :epkg:`scikit-learn` and :epkg:`cython`.
The first experiment implements a criterion for
a :epkg:`sklearn:tree:DecisionTreeRegressor`. This
code is based on the API in
`Criterion <https://github.com/scikit-learn/scikit-learn/blob/master/sklearn/tree/_criterion.pxd#L56>`_
which changed in version 0.21.



.. toctree::
    :maxdepth: 1

    notebooks/piecewise_linear_regression_criterion

.. raw:: html

    <div class="sphx-pyq-thumb" tooltip="A DecisionTreeRegressor can be trained with a couple of possible criterions but it is possible to implement a custom one (see hellinger_distance_criterion. See also tutorial Cython example of exposing C-computed arrays in Python without data copies which describes a way to implement fast cython extensions.">

.. only:: html

    .. figure:: /notebooks/piecewise_linear_regression_criterion.thumb.png

        :ref:`piecewiselinearregressioncriterionrst`

.. raw:: html

    </div>

.. raw:: html

   <div style='clear:both'></div>


Games with (scikit-learn) trees
===============================

The notebooks explore trees, mostly trees from :epkg:`scikit-learn`,
and compute unusual results from the structure.




.. toctree::
    :maxdepth: 1

    notebooks/leave_neighbors

.. raw:: html

    <div class="sphx-pyq-thumb" tooltip="A decision tree computes a partition of the feature space. We can wonder which leave is close to another one even though the predict the same value (or class). Do they share a border ?">

.. only:: html

    .. figure:: /notebooks/leave_neighbors.thumb.png

        :ref:`leaveneighborsrst`

.. raw:: html

    </div>




.. toctree::
    :hidden: 

    all_notebooks_coverage