Python sklearn.svm.NuSVR() Examples

def test_gradient_boosting_with_init_pipeline():
    # Check that the init estimator can be a pipeline (see issue #13466)

    X, y = make_regression(random_state=0)
    init = make_pipeline(LinearRegression())
    gb = GradientBoostingRegressor(init=init)
    gb.fit(X, y)  # pipeline without sample_weight works fine

    with pytest.raises(
            ValueError,
            match='The initial estimator Pipeline does not support sample '
                  'weights'):
        gb.fit(X, y, sample_weight=np.ones(X.shape[0]))

    # Passing sample_weight to a pipeline raises a ValueError. This test makes
    # sure we make the distinction between ValueError raised by a pipeline that
    # was passed sample_weight, and a ValueError raised by a regular estimator
    # whose input checking failed.
    with pytest.raises(
            ValueError,
            match='nu <= 0 or nu > 1'):
        # Note that NuSVR properly supports sample_weight
        init = NuSVR(gamma='auto', nu=1.5)
        gb = GradientBoostingRegressor(init=init)
        gb.fit(X, y, sample_weight=np.ones(X.shape[0])) 

def test_sample_weights():
    # Test weights on individual samples
    # TODO: check on NuSVR, OneClass, etc.
    clf = svm.SVC(gamma="scale")
    clf.fit(X, Y)
    assert_array_equal(clf.predict([X[2]]), [1.])

    sample_weight = [.1] * 3 + [10] * 3
    clf.fit(X, Y, sample_weight=sample_weight)
    assert_array_equal(clf.predict([X[2]]), [2.])

    # test that rescaling all samples is the same as changing C
    clf = svm.SVC(gamma="scale")
    clf.fit(X, Y)
    dual_coef_no_weight = clf.dual_coef_
    clf.set_params(C=100)
    clf.fit(X, Y, sample_weight=np.repeat(0.01, len(X)))
    assert_array_almost_equal(dual_coef_no_weight, clf.dual_coef_) 

def test_sample_weights():
    # Test weights on individual samples
    # TODO: check on NuSVR, OneClass, etc.
    clf = svm.SVC()
    clf.fit(X, Y)
    assert_array_equal(clf.predict([X[2]]), [1.])

    sample_weight = [.1] * 3 + [10] * 3
    clf.fit(X, Y, sample_weight=sample_weight)
    assert_array_equal(clf.predict([X[2]]), [2.])

    # test that rescaling all samples is the same as changing C
    clf = svm.SVC()
    clf.fit(X, Y)
    dual_coef_no_weight = clf.dual_coef_
    clf.set_params(C=100)
    clf.fit(X, Y, sample_weight=np.repeat(0.01, len(X)))
    assert_array_almost_equal(dual_coef_no_weight, clf.dual_coef_) 

def convert(model, feature_names, target):
    """Convert a Nu Support Vector Regression (NuSVR) model to the protobuf spec.
    Parameters
    ----------
    model: NuSVR
        A trained NuSVR encoder model.

    feature_names: [str]
        Name of the input columns.

    target: str
        Name of the output column.

    Returns
    -------
    model_spec: An object of type Model_pb.
        Protobuf representation of the model
    """
    if not (_HAS_SKLEARN):
        raise RuntimeError(
            "scikit-learn not found. scikit-learn conversion API is disabled."
        )

    _sklearn_util.check_expected_type(model, _NuSVR)
    return _SVR.convert(model, feature_names, target) 

def test_svr():
    # Test Support Vector Regression

    diabetes = datasets.load_diabetes()
    for clf in (svm.NuSVR(kernel='linear', nu=.4, C=1.0),
                svm.NuSVR(kernel='linear', nu=.4, C=10.),
                svm.SVR(kernel='linear', C=10.),
                svm.LinearSVR(C=10.),
                svm.LinearSVR(C=10.),
                ):
        clf.fit(diabetes.data, diabetes.target)
        assert_greater(clf.score(diabetes.data, diabetes.target), 0.02)

    # non-regression test; previously, BaseLibSVM would check that
    # len(np.unique(y)) < 2, which must only be done for SVC
    svm.SVR().fit(diabetes.data, np.ones(len(diabetes.data)))
    svm.LinearSVR().fit(diabetes.data, np.ones(len(diabetes.data))) 

def test_convert_nusvr_bool(self):
        model, X = fit_regression_model(
            NuSVR(), is_bool=True)
        model_onnx = convert_sklearn(
            model,
            "NuSVR",
            [("input", BooleanTensorType([None, X.shape[1]]))],
        )
        self.assertIsNotNone(model_onnx)
        dump_data_and_model(
            X,
            model,
            model_onnx,
            basename="SklearnNuSVRBool",
            allow_failure="StrictVersion(onnxruntime.__version__)"
                          " <= StrictVersion('0.2.1')"
        ) 

def test_convert_nusvr(self):
        model, X = self._fit_binary_classification(NuSVR())
        model_onnx = convert_sklearn(
            model, "SVR", [("input", FloatTensorType([None, X.shape[1]]))])
        node = model_onnx.graph.node[0]
        self.assertIsNotNone(node)
        self._check_attributes(
            node,
            {
                "coefficients": None,
                "kernel_params": None,
                "kernel_type": "RBF",
                "post_transform": None,
                "rho": None,
                "support_vectors": None,
            },
        )
        dump_data_and_model(X, model, model_onnx,
                            basename="SklearnRegNuSVR") 

def test_convert_nusvr_int(self):
        model, X = fit_regression_model(
            NuSVR(), is_int=True)
        model_onnx = convert_sklearn(
            model,
            "NuSVR",
            [("input", Int64TensorType([None, X.shape[1]]))],
        )
        self.assertIsNotNone(model_onnx)
        dump_data_and_model(
            X,
            model,
            model_onnx,
            basename="SklearnNuSVRInt-Dec4",
            allow_failure="StrictVersion(onnxruntime.__version__)"
                          " <= StrictVersion('0.2.1')"
        ) 

def test_svr_coef_sign():
    # Test that SVR(kernel="linear") has coef_ with the right sign.
    # Non-regression test for #2933.
    X = np.random.RandomState(21).randn(10, 3)
    y = np.random.RandomState(12).randn(10)

    for svr in [svm.SVR(kernel='linear'), svm.NuSVR(kernel='linear'),
                svm.LinearSVR()]:
        svr.fit(X, y)
        assert_array_almost_equal(svr.predict(X),
                                  np.dot(X, svr.coef_.ravel()) + svr.intercept_) 

def test_unfitted():
    X = "foo!"  # input validation not required when SVM not fitted

    clf = svm.SVC()
    assert_raises_regexp(Exception, r".*\bSVC\b.*\bnot\b.*\bfitted\b",
                         clf.predict, X)

    clf = svm.NuSVR()
    assert_raises_regexp(Exception, r".*\bNuSVR\b.*\bnot\b.*\bfitted\b",
                         clf.predict, X)


# ignore convergence warnings from max_iter=1 

def test_immutable_coef_property():
    # Check that primal coef modification are not silently ignored
    svms = [
        svm.SVC(kernel='linear').fit(iris.data, iris.target),
        svm.NuSVC(kernel='linear').fit(iris.data, iris.target),
        svm.SVR(kernel='linear').fit(iris.data, iris.target),
        svm.NuSVR(kernel='linear').fit(iris.data, iris.target),
        svm.OneClassSVM(kernel='linear').fit(iris.data),
    ]
    for clf in svms:
        assert_raises(AttributeError, clf.__setattr__, 'coef_', np.arange(3))
        assert_raises((RuntimeError, ValueError),
                      clf.coef_.__setitem__, (0, 0), 0) 

def test_objectmapper(self):
        df = pdml.ModelFrame([])
        self.assertIs(df.svm.SVC, svm.SVC)
        self.assertIs(df.svm.LinearSVC, svm.LinearSVC)
        self.assertIs(df.svm.NuSVC, svm.NuSVC)
        self.assertIs(df.svm.SVR, svm.SVR)
        self.assertIs(df.svm.NuSVR, svm.NuSVR)
        self.assertIs(df.svm.OneClassSVM, svm.OneClassSVM) 

def test_convert_nusvr_default(self):
        model, X = self._fit_binary_classification(NuSVR())
        model_onnx = convert_sklearn(
            model, "SVR", [("input", FloatTensorType([None, X.shape[1]]))])
        self.assertIsNotNone(model_onnx)
        dump_data_and_model(X, model, model_onnx, basename="SklearnRegNuSVR2") 

def test_conversion_bad_inputs(self):
        # Error on converting an untrained model
        with self.assertRaises(TypeError):
            model = NuSVR()
            spec = scikit_converter.convert(model, "data", "out")

        # Check the expected class during covnersion.
        with self.assertRaises(TypeError):
            model = OneHotEncoder()
            spec = scikit_converter.convert(model, "data", "out") 

def setUpClass(self):
        """
        Set up the unit test by loading the dataset and training a model.
        """
        if not _HAS_SKLEARN:
            return

        self.scikit_model = NuSVR(kernel="linear")
        self.data = load_boston()
        self.scikit_model.fit(self.data["data"], self.data["target"]) 

def test_unfitted():
    X = "foo!"  # input validation not required when SVM not fitted

    clf = svm.SVC(gamma="scale")
    assert_raises_regexp(Exception, r".*\bSVC\b.*\bnot\b.*\bfitted\b",
                         clf.predict, X)

    clf = svm.NuSVR(gamma='scale')
    assert_raises_regexp(Exception, r".*\bNuSVR\b.*\bnot\b.*\bfitted\b",
                         clf.predict, X)


# ignore convergence warnings from max_iter=1 

def test_immutable_coef_property():
    # Check that primal coef modification are not silently ignored
    svms = [
        svm.SVC(kernel='linear').fit(iris.data, iris.target),
        svm.NuSVC(kernel='linear').fit(iris.data, iris.target),
        svm.SVR(kernel='linear').fit(iris.data, iris.target),
        svm.NuSVR(kernel='linear').fit(iris.data, iris.target),
        svm.OneClassSVM(kernel='linear').fit(iris.data),
    ]
    for clf in svms:
        assert_raises(AttributeError, clf.__setattr__, 'coef_', np.arange(3))
        assert_raises((RuntimeError, ValueError),
                      clf.coef_.__setitem__, (0, 0), 0) 

def test_svr():
    # Test Support Vector Regression

    diabetes = datasets.load_diabetes()
    for clf in (svm.NuSVR(kernel='linear', nu=.4, C=1.0),
                svm.NuSVR(kernel='linear', nu=.4, C=10.),
                svm.SVR(kernel='linear', C=10.),
                svm.LinearSVR(C=10.),
                svm.LinearSVR(C=10.),
                ):
        clf.fit(diabetes.data, diabetes.target)
        assert_greater(clf.score(diabetes.data, diabetes.target), 0.02)

    # non-regression test; previously, BaseLibSVM would check that
    # len(np.unique(y)) < 2, which must only be done for SVC
    svm.SVR(gamma='scale').fit(diabetes.data, np.ones(len(diabetes.data)))
    svm.LinearSVR().fit(diabetes.data, np.ones(len(diabetes.data))) 

def _test_evaluation(self, allow_slow):
        """
        Test that the same predictions are made
        """

        # Generate some smallish (some kernels take too long on anything else) random data
        x, y = [], []
        for _ in range(50):
            cur_x1, cur_x2 = random.gauss(2, 3), random.gauss(-1, 2)
            x.append([cur_x1, cur_x2])
            y.append(1 + 2 * cur_x1 + 3 * cur_x2)

        input_names = ["x1", "x2"]
        df = pd.DataFrame(x, columns=input_names)

        # Parameters to test
        kernel_parameters = [
            {},
            {"kernel": "rbf", "gamma": 1.2},
            {"kernel": "linear"},
            {"kernel": "poly"},
            {"kernel": "poly", "degree": 2},
            {"kernel": "poly", "gamma": 0.75},
            {"kernel": "poly", "degree": 0, "gamma": 0.9, "coef0": 2},
            {"kernel": "sigmoid"},
            {"kernel": "sigmoid", "gamma": 1.3},
            {"kernel": "sigmoid", "coef0": 0.8},
            {"kernel": "sigmoid", "coef0": 0.8, "gamma": 0.5},
        ]
        non_kernel_parameters = [
            {},
            {"C": 1},
            {"C": 1.5, "shrinking": True},
            {"C": 0.5, "shrinking": False, "nu": 0.9},
        ]

        # Test
        for param1 in non_kernel_parameters:
            for param2 in kernel_parameters:
                cur_params = param1.copy()
                cur_params.update(param2)

                cur_model = NuSVR(**cur_params)
                cur_model.fit(x, y)
                df["prediction"] = cur_model.predict(x)

                spec = scikit_converter.convert(cur_model, input_names, "target")

                if _is_macos() and _macos_version() >= (10, 13):
                    metrics = evaluate_regressor(spec, df)
                    self.assertAlmostEquals(metrics["max_error"], 0)

                if not allow_slow:
                    break

            if not allow_slow:
                break 

def fit(self, X, y, sample_weight=None, eval_set=None, sample_weight_eval_set=None, **kwargs):
        X = dt.Frame(X)

        orig_cols = list(X.names)

        if self.num_classes >= 2:
            feature_model = NuSVC(kernel='linear', nu=self.params['nu'])
            model = NuSVC(nu=self.params['nu'], kernel=self.params['kernel'],
                          degree=self.params['degree'], probability=self.params['probability'])

            lb = LabelEncoder()
            lb.fit(self.labels)
            y = lb.transform(y)
        else:
            feature_model = NuSVR(kernel='linear', nu=self.params['nu'])
            model = NuSVR(nu=self.params['nu'], kernel=self.params['kernel'],
                          degree=self.params['degree'])

        self.means = dict()

        for col in X.names:
            XX = X[:, col]
            self.means[col] = XX.mean1()
            if self.means[col] is None:
                self.means[col] = 0
            XX.replace(None, self.means[col])
            X[:, col] = XX
            assert X[dt.isna(dt.f[col]), col].nrows == 0

        X = X.to_numpy()

        # nu is infeasible sometimes
        # doing quaternary search on both sides of selected nu
        valid_nu = None
        while valid_nu is None:
            try:
                model.fit(X, y)
                valid_nu = self.params['nu']
            except:
                if self.params['nu'] > 0.5:
                    self.params['nu'] = 1.0 - self.params['nu']
                else:
                    self.params['nu'] = (4.0 - 3.0 * self.params['nu']) / 4.0
                if self.num_classes >= 2:
                    feature_model = NuSVC(kernel='linear', nu=self.params['nu'])
                    model = NuSVC(nu=self.params['nu'], kernel=self.params['kernel'],
                                  degree=self.params['degree'], probability=self.params['probability'])
                else:
                    feature_model = NuSVR(kernel='linear', nu=self.params['nu'])
                    model = NuSVR(nu=self.params['nu'], kernel=self.params['kernel'],
                                  degree=self.params['degree'])

        feature_model.fit(X, y)
        importances = np.array(abs(feature_model.coef_)).ravel()

        self.set_model_properties(model=model,
                                  features=orig_cols,
                                  importances=importances.tolist(),
                                  iterations=0)