.. _2018-09-18sklearnapirst:

================================
2018-09-18 - API de scikit-learn
================================


.. only:: html

    **Links:** :download:`notebook <2018-09-18_sklearn_api.ipynb>`, :downloadlink:`html <2018-09-18_sklearn_api2html.html>`, :download:`python <2018-09-18_sklearn_api.py>`, :downloadlink:`slides <2018-09-18_sklearn_api.slides.html>`, :githublink:`GitHub|_doc/notebooks/notebook_eleves/2018-2019/2018-09-18_sklearn_api.ipynb|*`


Présentation de l’API de *scikit-learn* et implémentation d’un
prédicteur fait maison. On utilise le jeu du Titanic qu’on peut
récupérer sur
`opendatasoft <https://public.opendatasoft.com/explore/dataset/titanic-passengers/?flg=fr>`__
ou
`awesome-public-datasets <https://github.com/awesomedata/awesome-public-datasets/tree/master/Datasets>`__.

.. code:: ipython3

    import pandas
    df = pandas.read_csv("titanic.csv/titanic.csv")
    df.head(n=2)






.. raw:: html

    <div>
    <style scoped>
        .dataframe tbody tr th:only-of-type {
            vertical-align: middle;
        }

        .dataframe tbody tr th {
            vertical-align: top;
        }

        .dataframe thead th {
            text-align: right;
        }
    </style>
    <table border="1" class="dataframe">
      <thead>
        <tr style="text-align: right;">
          <th></th>
          <th>PassengerId</th>
          <th>Survived</th>
          <th>Pclass</th>
          <th>Name</th>
          <th>Sex</th>
          <th>Age</th>
          <th>SibSp</th>
          <th>Parch</th>
          <th>Ticket</th>
          <th>Fare</th>
          <th>Cabin</th>
          <th>Embarked</th>
        </tr>
      </thead>
      <tbody>
        <tr>
          <th>0</th>
          <td>1</td>
          <td>0</td>
          <td>3</td>
          <td>Braund, Mr. Owen Harris</td>
          <td>male</td>
          <td>22.0</td>
          <td>1</td>
          <td>0</td>
          <td>A/5 21171</td>
          <td>7.2500</td>
          <td>NaN</td>
          <td>S</td>
        </tr>
        <tr>
          <th>1</th>
          <td>2</td>
          <td>1</td>
          <td>1</td>
          <td>Cumings, Mrs. John Bradley (Florence Briggs Th...</td>
          <td>female</td>
          <td>38.0</td>
          <td>1</td>
          <td>0</td>
          <td>PC 17599</td>
          <td>71.2833</td>
          <td>C85</td>
          <td>C</td>
        </tr>
      </tbody>
    </table>
    </div>



.. code:: ipython3

    X, y = df[["Age", "Fare"]], df['Survived']

.. code:: ipython3

    from sklearn.model_selection import train_test_split
    X_train, X_test, y_train, y_test = train_test_split(X, y)

.. code:: ipython3

    from sklearn.linear_model import LogisticRegression
    cls = LogisticRegression()
    try:
        cls.fit(X_train, y_train)
    except Exception as e:
        print(e)


.. parsed-literal::
    Input contains NaN, infinity or a value too large for dtype('float64').


.. code:: ipython3

    try:
        from sklearn.impute import SimpleImputer as Imputer
    except ImportError:
        from sklearn.preprocessing import Imputer
    imp = Imputer()
    imp.fit(X_train)
    X_train_nomiss = imp.transform(X_train)

.. code:: ipython3

    cls = LogisticRegression()
    cls.fit(X_train_nomiss, y_train)




.. parsed-literal::
    LogisticRegression(C=1.0, class_weight=None, dual=False, fit_intercept=True,
                       intercept_scaling=1, l1_ratio=None, max_iter=100,
                       multi_class='auto', n_jobs=None, penalty='l2',
                       random_state=None, solver='lbfgs', tol=0.0001, verbose=0,
                       warm_start=False)



.. code:: ipython3

    cls.score(imp.transform(X_test), y_test)




.. parsed-literal::
    0.6502242152466368



.. code:: ipython3

    from sklearn.pipeline import Pipeline
    pipe = Pipeline([("imputer", Imputer()), 
                     ("lr", LogisticRegression())])
    pipe.fit(X_train, y_train)




.. parsed-literal::
    Pipeline(memory=None,
             steps=[('imputer',
                     SimpleImputer(add_indicator=False, copy=True, fill_value=None,
                                   missing_values=nan, strategy='mean',
                                   verbose=0)),
                    ('lr',
                     LogisticRegression(C=1.0, class_weight=None, dual=False,
                                        fit_intercept=True, intercept_scaling=1,
                                        l1_ratio=None, max_iter=100,
                                        multi_class='auto', n_jobs=None,
                                        penalty='l2', random_state=None,
                                        solver='lbfgs', tol=0.0001, verbose=0,
                                        warm_start=False))],
             verbose=False)



.. code:: ipython3

    pipe.score(X_test, y_test)




.. parsed-literal::
    0.6502242152466368



.. code:: ipython3

    from sklearn.model_selection import GridSearchCV
    grid = GridSearchCV(pipe, {"imputer__strategy": ['mean', 'most_frequent'],
                               "lr__max_iter": [5, 10, 50]})
    grid.fit(X_train, y_train)


.. parsed-literal::
    C:\xavierdupre\__home_\github_fork\scikit-learn\sklearn\linear_model\logistic.py:935: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.
      "of iterations.", ConvergenceWarning)
    C:\xavierdupre\__home_\github_fork\scikit-learn\sklearn\linear_model\logistic.py:935: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.
      "of iterations.", ConvergenceWarning)
    C:\xavierdupre\__home_\github_fork\scikit-learn\sklearn\linear_model\logistic.py:935: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.
      "of iterations.", ConvergenceWarning)
    C:\xavierdupre\__home_\github_fork\scikit-learn\sklearn\linear_model\logistic.py:935: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.
      "of iterations.", ConvergenceWarning)
    C:\xavierdupre\__home_\github_fork\scikit-learn\sklearn\linear_model\logistic.py:935: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.
      "of iterations.", ConvergenceWarning)
    C:\xavierdupre\__home_\github_fork\scikit-learn\sklearn\linear_model\logistic.py:935: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.
      "of iterations.", ConvergenceWarning)
    C:\xavierdupre\__home_\github_fork\scikit-learn\sklearn\linear_model\logistic.py:935: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.
      "of iterations.", ConvergenceWarning)
    C:\xavierdupre\__home_\github_fork\scikit-learn\sklearn\linear_model\logistic.py:935: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.
      "of iterations.", ConvergenceWarning)
    C:\xavierdupre\__home_\github_fork\scikit-learn\sklearn\linear_model\logistic.py:935: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.
      "of iterations.", ConvergenceWarning)
    C:\xavierdupre\__home_\github_fork\scikit-learn\sklearn\linear_model\logistic.py:935: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.
      "of iterations.", ConvergenceWarning)
    C:\xavierdupre\__home_\github_fork\scikit-learn\sklearn\linear_model\logistic.py:935: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.
      "of iterations.", ConvergenceWarning)
    C:\xavierdupre\__home_\github_fork\scikit-learn\sklearn\linear_model\logistic.py:935: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.
      "of iterations.", ConvergenceWarning)
    C:\xavierdupre\__home_\github_fork\scikit-learn\sklearn\linear_model\logistic.py:935: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.
      "of iterations.", ConvergenceWarning)
    C:\xavierdupre\__home_\github_fork\scikit-learn\sklearn\linear_model\logistic.py:935: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.
      "of iterations.", ConvergenceWarning)
    C:\xavierdupre\__home_\github_fork\scikit-learn\sklearn\linear_model\logistic.py:935: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.
      "of iterations.", ConvergenceWarning)
    C:\xavierdupre\__home_\github_fork\scikit-learn\sklearn\linear_model\logistic.py:935: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.
      "of iterations.", ConvergenceWarning)
    C:\xavierdupre\__home_\github_fork\scikit-learn\sklearn\linear_model\logistic.py:935: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.
      "of iterations.", ConvergenceWarning)
    C:\xavierdupre\__home_\github_fork\scikit-learn\sklearn\linear_model\logistic.py:935: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.
      "of iterations.", ConvergenceWarning)
    C:\xavierdupre\__home_\github_fork\scikit-learn\sklearn\linear_model\logistic.py:935: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.
      "of iterations.", ConvergenceWarning)
    C:\xavierdupre\__home_\github_fork\scikit-learn\sklearn\linear_model\logistic.py:935: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.
      "of iterations.", ConvergenceWarning)
    C:\xavierdupre\__home_\github_fork\scikit-learn\sklearn\linear_model\logistic.py:935: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.
      "of iterations.", ConvergenceWarning)




.. parsed-literal::

    GridSearchCV(cv=None, error_score=nan,
                 estimator=Pipeline(memory=None,
                                    steps=[('imputer',
                                            SimpleImputer(add_indicator=False,
                                                          copy=True,
                                                          fill_value=None,
                                                          missing_values=nan,
                                                          strategy='mean',
                                                          verbose=0)),
                                           ('lr',
                                            LogisticRegression(C=1.0,
                                                               class_weight=None,
                                                               dual=False,
                                                               fit_intercept=True,
                                                               intercept_scaling=1,
                                                               l1_ratio=None,
                                                               max_iter=100,
                                                               multi_class='auto',
                                                               n_jobs=None,
                                                               penalty='l2',
                                                               random_state=None,
                                                               solver='lbfgs',
                                                               tol=0.0001,
                                                               verbose=0,
                                                               warm_start=False))],
                                    verbose=False),
                 iid='deprecated', n_jobs=None,
                 param_grid={'imputer__strategy': ['mean', 'most_frequent'],
                             'lr__max_iter': [5, 10, 50]},
                 pre_dispatch='2*n_jobs', refit=True, return_train_score=False,
                 scoring=None, verbose=0)



.. code:: ipython3

    res = pandas.DataFrame(grid.cv_results_)
    col = [_ for _ in res.columns if 'param_' in _ or "test_score" in _]
    res[col].T






.. raw:: html

    <div>
    <style scoped>
        .dataframe tbody tr th:only-of-type {
            vertical-align: middle;
        }

        .dataframe tbody tr th {
            vertical-align: top;
        }

        .dataframe thead th {
            text-align: right;
        }
    </style>
    <table border="1" class="dataframe">
      <thead>
        <tr style="text-align: right;">
          <th></th>
          <th>0</th>
          <th>1</th>
          <th>2</th>
          <th>3</th>
          <th>4</th>
          <th>5</th>
        </tr>
      </thead>
      <tbody>
        <tr>
          <th>param_imputer__strategy</th>
          <td>mean</td>
          <td>mean</td>
          <td>mean</td>
          <td>most_frequent</td>
          <td>most_frequent</td>
          <td>most_frequent</td>
        </tr>
        <tr>
          <th>param_lr__max_iter</th>
          <td>5</td>
          <td>10</td>
          <td>50</td>
          <td>5</td>
          <td>10</td>
          <td>50</td>
        </tr>
        <tr>
          <th>split0_test_score</th>
          <td>0.686567</td>
          <td>0.69403</td>
          <td>0.656716</td>
          <td>0.686567</td>
          <td>0.69403</td>
          <td>0.656716</td>
        </tr>
        <tr>
          <th>split1_test_score</th>
          <td>0.619403</td>
          <td>0.604478</td>
          <td>0.597015</td>
          <td>0.61194</td>
          <td>0.626866</td>
          <td>0.61194</td>
        </tr>
        <tr>
          <th>split2_test_score</th>
          <td>0.679104</td>
          <td>0.679104</td>
          <td>0.671642</td>
          <td>0.664179</td>
          <td>0.671642</td>
          <td>0.656716</td>
        </tr>
        <tr>
          <th>split3_test_score</th>
          <td>0.706767</td>
          <td>0.699248</td>
          <td>0.684211</td>
          <td>0.706767</td>
          <td>0.714286</td>
          <td>0.684211</td>
        </tr>
        <tr>
          <th>split4_test_score</th>
          <td>0.676692</td>
          <td>0.699248</td>
          <td>0.699248</td>
          <td>0.676692</td>
          <td>0.706767</td>
          <td>0.691729</td>
        </tr>
        <tr>
          <th>mean_test_score</th>
          <td>0.673707</td>
          <td>0.675222</td>
          <td>0.661766</td>
          <td>0.669229</td>
          <td>0.682718</td>
          <td>0.660263</td>
        </tr>
        <tr>
          <th>std_test_score</th>
          <td>0.0291387</td>
          <td>0.0361333</td>
          <td>0.0352828</td>
          <td>0.0318525</td>
          <td>0.0314484</td>
          <td>0.0280138</td>
        </tr>
        <tr>
          <th>rank_test_score</th>
          <td>3</td>
          <td>2</td>
          <td>5</td>
          <td>4</td>
          <td>1</td>
          <td>6</td>
        </tr>
      </tbody>
    </table>
    </div>



.. code:: ipython3

    from sklearn.base import BaseEstimator, ClassifierMixin
    import numpy
    
    class MeanPredictor(BaseEstimator, ClassifierMixin):
        def __init__(self, alpha=0.5):
            self.alpha = alpha
            
        def fit(self, X, y):
            self.mean_ = int(self.alpha + numpy.mean(y))
            
        def predict(self, X):
            return numpy.array(list(self.mean_ for k in range(X.shape[0])))

.. code:: ipython3

    pipe_mean = Pipeline([('imputer', Imputer()), 
                          ('meanpredictor', MeanPredictor())])
    pipe_mean.fit(X_train, y_train)




.. parsed-literal::
    Pipeline(memory=None,
             steps=[('imputer',
                     SimpleImputer(add_indicator=False, copy=True, fill_value=None,
                                   missing_values=nan, strategy='mean',
                                   verbose=0)),
                    ('meanpredictor', MeanPredictor(alpha=0.5))],
             verbose=False)



.. code:: ipython3

    pipe_mean.score(X_test, y_test)




.. parsed-literal::
    0.6322869955156951



.. code:: ipython3

    from sklearn.model_selection import GridSearchCV
    grid = GridSearchCV(pipe_mean, {"imputer__strategy": ['mean', 'most_frequent'],
                                    "meanpredictor__alpha": [0.2, 0.5, 0.8]})
    grid.fit(X_train, y_train)




.. parsed-literal::
    GridSearchCV(cv=None, error_score=nan,
                 estimator=Pipeline(memory=None,
                                    steps=[('imputer',
                                            SimpleImputer(add_indicator=False,
                                                          copy=True,
                                                          fill_value=None,
                                                          missing_values=nan,
                                                          strategy='mean',
                                                          verbose=0)),
                                           ('meanpredictor',
                                            MeanPredictor(alpha=0.5))],
                                    verbose=False),
                 iid='deprecated', n_jobs=None,
                 param_grid={'imputer__strategy': ['mean', 'most_frequent'],
                             'meanpredictor__alpha': [0.2, 0.5, 0.8]},
                 pre_dispatch='2*n_jobs', refit=True, return_train_score=False,
                 scoring=None, verbose=0)



.. code:: ipython3

    res = pandas.DataFrame(grid.cv_results_)
    col = [_ for _ in res.columns if 'param_' in _ or "test_score" in _]
    res[col].T






.. raw:: html

    <div>
    <style scoped>
        .dataframe tbody tr th:only-of-type {
            vertical-align: middle;
        }

        .dataframe tbody tr th {
            vertical-align: top;
        }

        .dataframe thead th {
            text-align: right;
        }
    </style>
    <table border="1" class="dataframe">
      <thead>
        <tr style="text-align: right;">
          <th></th>
          <th>0</th>
          <th>1</th>
          <th>2</th>
          <th>3</th>
          <th>4</th>
          <th>5</th>
        </tr>
      </thead>
      <tbody>
        <tr>
          <th>param_imputer__strategy</th>
          <td>mean</td>
          <td>mean</td>
          <td>mean</td>
          <td>most_frequent</td>
          <td>most_frequent</td>
          <td>most_frequent</td>
        </tr>
        <tr>
          <th>param_meanpredictor__alpha</th>
          <td>0.2</td>
          <td>0.5</td>
          <td>0.8</td>
          <td>0.2</td>
          <td>0.5</td>
          <td>0.8</td>
        </tr>
        <tr>
          <th>split0_test_score</th>
          <td>0.61194</td>
          <td>0.61194</td>
          <td>0.38806</td>
          <td>0.61194</td>
          <td>0.61194</td>
          <td>0.38806</td>
        </tr>
        <tr>
          <th>split1_test_score</th>
          <td>0.61194</td>
          <td>0.61194</td>
          <td>0.38806</td>
          <td>0.61194</td>
          <td>0.61194</td>
          <td>0.38806</td>
        </tr>
        <tr>
          <th>split2_test_score</th>
          <td>0.61194</td>
          <td>0.61194</td>
          <td>0.38806</td>
          <td>0.61194</td>
          <td>0.61194</td>
          <td>0.38806</td>
        </tr>
        <tr>
          <th>split3_test_score</th>
          <td>0.609023</td>
          <td>0.609023</td>
          <td>0.390977</td>
          <td>0.609023</td>
          <td>0.609023</td>
          <td>0.390977</td>
        </tr>
        <tr>
          <th>split4_test_score</th>
          <td>0.609023</td>
          <td>0.609023</td>
          <td>0.390977</td>
          <td>0.609023</td>
          <td>0.609023</td>
          <td>0.390977</td>
        </tr>
        <tr>
          <th>mean_test_score</th>
          <td>0.610773</td>
          <td>0.610773</td>
          <td>0.389227</td>
          <td>0.610773</td>
          <td>0.610773</td>
          <td>0.389227</td>
        </tr>
        <tr>
          <th>std_test_score</th>
          <td>0.0014294</td>
          <td>0.0014294</td>
          <td>0.0014294</td>
          <td>0.0014294</td>
          <td>0.0014294</td>
          <td>0.0014294</td>
        </tr>
        <tr>
          <th>rank_test_score</th>
          <td>1</td>
          <td>1</td>
          <td>5</td>
          <td>1</td>
          <td>1</td>
          <td>5</td>
        </tr>
      </tbody>
    </table>
    </div>



.. code:: ipython3

    best = grid.best_estimator_

.. code:: ipython3

    import pickle
    with open("model.pkl", "wb") as f:
        pickle.dump(best, f)

.. code:: ipython3

    with open("model.pkl", "rb") as f:
        model = pickle.load(f)

.. code:: ipython3

    model.predict(X_test) == best.predict(X_test)




.. parsed-literal::
    array([ True,  True,  True,  True,  True,  True,  True,  True,  True,
            True,  True,  True,  True,  True,  True,  True,  True,  True,
            True,  True,  True,  True,  True,  True,  True,  True,  True,
            True,  True,  True,  True,  True,  True,  True,  True,  True,
            True,  True,  True,  True,  True,  True,  True,  True,  True,
            True,  True,  True,  True,  True,  True,  True,  True,  True,
            True,  True,  True,  True,  True,  True,  True,  True,  True,
            True,  True,  True,  True,  True,  True,  True,  True,  True,
            True,  True,  True,  True,  True,  True,  True,  True,  True,
            True,  True,  True,  True,  True,  True,  True,  True,  True,
            True,  True,  True,  True,  True,  True,  True,  True,  True,
            True,  True,  True,  True,  True,  True,  True,  True,  True,
            True,  True,  True,  True,  True,  True,  True,  True,  True,
            True,  True,  True,  True,  True,  True,  True,  True,  True,
            True,  True,  True,  True,  True,  True,  True,  True,  True,
            True,  True,  True,  True,  True,  True,  True,  True,  True,
            True,  True,  True,  True,  True,  True,  True,  True,  True,
            True,  True,  True,  True,  True,  True,  True,  True,  True,
            True,  True,  True,  True,  True,  True,  True,  True,  True,
            True,  True,  True,  True,  True,  True,  True,  True,  True,
            True,  True,  True,  True,  True,  True,  True,  True,  True,
            True,  True,  True,  True,  True,  True,  True,  True,  True,
            True,  True,  True,  True,  True,  True,  True,  True,  True,
            True,  True,  True,  True,  True,  True,  True,  True,  True,
            True,  True,  True,  True,  True,  True,  True])