Python sklearn.model_selection.validation_curve() Examples

def test_validation_curve(self):
        digits = datasets.load_digits()
        df = pdml.ModelFrame(digits)

        param_range = np.logspace(-2, -1, 2)

        svc = df.svm.SVC(random_state=self.random_state)
        result = df.model_selection.validation_curve(svc, 'gamma',
                                                     param_range)
        expected = ms.validation_curve(svm.SVC(random_state=self.random_state),
                                       digits.data, digits.target,
                                       'gamma', param_range)

        self.assertEqual(len(result), 2)
        self.assert_numpy_array_almost_equal(result[0], expected[0])
        self.assert_numpy_array_almost_equal(result[1], expected[1]) 

def test_validation_curve():
    X, y = make_classification(n_samples=2, n_features=1, n_informative=1,
                               n_redundant=0, n_classes=2,
                               n_clusters_per_class=1, random_state=0)
    param_range = np.linspace(0, 1, 10)
    with warnings.catch_warnings(record=True) as w:
        train_scores, test_scores = validation_curve(
            MockEstimatorWithParameter(), X, y, param_name="param",
            param_range=param_range, cv=2
        )
    if len(w) > 0:
        raise RuntimeError("Unexpected warning: %r" % w[0].message)

    assert_array_almost_equal(train_scores.mean(axis=1), param_range)
    assert_array_almost_equal(test_scores.mean(axis=1), 1 - param_range) 

def test_validation_curve_clone_estimator():
    X, y = make_classification(n_samples=2, n_features=1, n_informative=1,
                               n_redundant=0, n_classes=2,
                               n_clusters_per_class=1, random_state=0)

    param_range = np.linspace(1, 0, 10)
    _, _ = validation_curve(
        MockEstimatorWithSingleFitCallAllowed(), X, y,
        param_name="param", param_range=param_range, cv=2
    ) 

def test_validation_curve_cv_splits_consistency():
    n_samples = 100
    n_splits = 5
    X, y = make_classification(n_samples=100, random_state=0)

    scores1 = validation_curve(SVC(kernel='linear', random_state=0), X, y,
                               'C', [0.1, 0.1, 0.2, 0.2],
                               cv=OneTimeSplitter(n_splits=n_splits,
                                                  n_samples=n_samples))
    # The OneTimeSplitter is a non-re-entrant cv splitter. Unless, the
    # `split` is called for each parameter, the following should produce
    # identical results for param setting 1 and param setting 2 as both have
    # the same C value.
    assert_array_almost_equal(*np.vsplit(np.hstack(scores1)[(0, 2, 1, 3), :],
                                         2))

    scores2 = validation_curve(SVC(kernel='linear', random_state=0), X, y,
                               'C', [0.1, 0.1, 0.2, 0.2],
                               cv=KFold(n_splits=n_splits, shuffle=True))

    # For scores2, compare the 1st and 2nd parameter's scores
    # (Since the C value for 1st two param setting is 0.1, they must be
    # consistent unless the train test folds differ between the param settings)
    assert_array_almost_equal(*np.vsplit(np.hstack(scores2)[(0, 2, 1, 3), :],
                                         2))

    scores3 = validation_curve(SVC(kernel='linear', random_state=0), X, y,
                               'C', [0.1, 0.1, 0.2, 0.2],
                               cv=KFold(n_splits=n_splits))

    # OneTimeSplitter is basically unshuffled KFold(n_splits=5). Sanity check.
    assert_array_almost_equal(np.array(scores3), np.array(scores1)) 

def test_validation_curve():
    X, y = make_classification(n_samples=2, n_features=1, n_informative=1,
                               n_redundant=0, n_classes=2,
                               n_clusters_per_class=1, random_state=0)
    param_range = np.linspace(0, 1, 10)
    with warnings.catch_warnings(record=True) as w:
        train_scores, test_scores = validation_curve(
            MockEstimatorWithParameter(), X, y, param_name="param",
            param_range=param_range, cv=2
        )
    if len(w) > 0:
        raise RuntimeError("Unexpected warning: %r" % w[0].message)

    assert_array_almost_equal(train_scores.mean(axis=1), param_range)
    assert_array_almost_equal(test_scores.mean(axis=1), 1 - param_range) 

def test_validation_curve_clone_estimator():
    X, y = make_classification(n_samples=2, n_features=1, n_informative=1,
                               n_redundant=0, n_classes=2,
                               n_clusters_per_class=1, random_state=0)

    param_range = np.linspace(1, 0, 10)
    _, _ = validation_curve(
        MockEstimatorWithSingleFitCallAllowed(), X, y,
        param_name="param", param_range=param_range, cv=2
    ) 

def test_validation_curve_cv_splits_consistency():
    n_samples = 100
    n_splits = 5
    X, y = make_classification(n_samples=100, random_state=0)

    scores1 = validation_curve(SVC(kernel='linear', random_state=0), X, y,
                               'C', [0.1, 0.1, 0.2, 0.2],
                               cv=OneTimeSplitter(n_splits=n_splits,
                                                  n_samples=n_samples))
    # The OneTimeSplitter is a non-re-entrant cv splitter. Unless, the
    # `split` is called for each parameter, the following should produce
    # identical results for param setting 1 and param setting 2 as both have
    # the same C value.
    assert_array_almost_equal(*np.vsplit(np.hstack(scores1)[(0, 2, 1, 3), :],
                                         2))

    scores2 = validation_curve(SVC(kernel='linear', random_state=0), X, y,
                               'C', [0.1, 0.1, 0.2, 0.2],
                               cv=KFold(n_splits=n_splits, shuffle=True))

    # For scores2, compare the 1st and 2nd parameter's scores
    # (Since the C value for 1st two param setting is 0.1, they must be
    # consistent unless the train test folds differ between the param settings)
    assert_array_almost_equal(*np.vsplit(np.hstack(scores2)[(0, 2, 1, 3), :],
                                         2))

    scores3 = validation_curve(SVC(kernel='linear', random_state=0), X, y,
                               'C', [0.1, 0.1, 0.2, 0.2],
                               cv=KFold(n_splits=n_splits))

    # OneTimeSplitter is basically unshuffled KFold(n_splits=5). Sanity check.
    assert_array_almost_equal(np.array(scores3), np.array(scores1)) 

def test_fit_and_score_failing():
    # Create a failing classifier to deliberately fail
    failing_clf = FailingClassifier(FailingClassifier.FAILING_PARAMETER)
    # dummy X data
    X = np.arange(1, 10)
    y = np.ones(9)
    fit_and_score_args = [failing_clf, X, None, dict(), None, None, 0,
                          None, None]
    # passing error score to trigger the warning message
    fit_and_score_kwargs = {'error_score': 0}
    # check if the warning message type is as expected
    assert_warns(FitFailedWarning, _fit_and_score, *fit_and_score_args,
                 **fit_and_score_kwargs)
    # since we're using FailingClassfier, our error will be the following
    error_message = "ValueError: Failing classifier failed as required"
    # the warning message we're expecting to see
    warning_message = ("Estimator fit failed. The score on this train-test "
                       "partition for these parameters will be set to %f. "
                       "Details: \n%s" % (fit_and_score_kwargs['error_score'],
                                          error_message))
    # check if the same warning is triggered
    assert_warns_message(FitFailedWarning, warning_message, _fit_and_score,
                         *fit_and_score_args, **fit_and_score_kwargs)

    # check if warning was raised, with default error_score argument
    warning_message = ("From version 0.22, errors during fit will result "
                       "in a cross validation score of NaN by default. Use "
                       "error_score='raise' if you want an exception "
                       "raised or error_score=np.nan to adopt the "
                       "behavior from version 0.22.")
    with pytest.raises(ValueError):
        assert_warns_message(FutureWarning, warning_message, _fit_and_score,
                             *fit_and_score_args)

    fit_and_score_kwargs = {'error_score': 'raise'}
    # check if exception was raised, with default error_score='raise'
    assert_raise_message(ValueError, "Failing classifier failed as required",
                         _fit_and_score, *fit_and_score_args,
                         **fit_and_score_kwargs)

    # check that functions upstream pass error_score param to _fit_and_score
    error_message = ("error_score must be the string 'raise' or a"
                     " numeric value. (Hint: if using 'raise', please"
                     " make sure that it has been spelled correctly.)")

    assert_raise_message(ValueError, error_message, cross_validate,
                         failing_clf, X, cv=3, error_score='unvalid-string')

    assert_raise_message(ValueError, error_message, cross_val_score,
                         failing_clf, X, cv=3, error_score='unvalid-string')

    assert_raise_message(ValueError, error_message, learning_curve,
                         failing_clf, X, y, cv=3, error_score='unvalid-string')

    assert_raise_message(ValueError, error_message, validation_curve,
                         failing_clf, X, y, 'parameter',
                         [FailingClassifier.FAILING_PARAMETER], cv=3,
                         error_score='unvalid-string')

    assert_equal(failing_clf.score(), 0.)  # FailingClassifier coverage