Python sklearn.multiclass.OneVsOneClassifier() Examples

def test_ovr_ovo_regressor():
    # test that ovr and ovo work on regressors which don't have a decision_
    # function
    ovr = OneVsRestClassifier(DecisionTreeRegressor())
    pred = ovr.fit(iris.data, iris.target).predict(iris.data)
    assert_equal(len(ovr.estimators_), n_classes)
    assert_array_equal(np.unique(pred), [0, 1, 2])
    # we are doing something sensible
    assert_greater(np.mean(pred == iris.target), .9)

    ovr = OneVsOneClassifier(DecisionTreeRegressor())
    pred = ovr.fit(iris.data, iris.target).predict(iris.data)
    assert_equal(len(ovr.estimators_), n_classes * (n_classes - 1) / 2)
    assert_array_equal(np.unique(pred), [0, 1, 2])
    # we are doing something sensible
    assert_greater(np.mean(pred == iris.target), .9) 

def test_ovr_ovo_regressor():
    # test that ovr and ovo work on regressors which don't have a decision_
    # function
    ovr = OneVsRestClassifier(DecisionTreeRegressor())
    pred = ovr.fit(iris.data, iris.target).predict(iris.data)
    assert_equal(len(ovr.estimators_), n_classes)
    assert_array_equal(np.unique(pred), [0, 1, 2])
    # we are doing something sensible
    assert_greater(np.mean(pred == iris.target), .9)

    ovr = OneVsOneClassifier(DecisionTreeRegressor())
    pred = ovr.fit(iris.data, iris.target).predict(iris.data)
    assert_equal(len(ovr.estimators_), n_classes * (n_classes - 1) / 2)
    assert_array_equal(np.unique(pred), [0, 1, 2])
    # we are doing something sensible
    assert_greater(np.mean(pred == iris.target), .9) 

def test_ovo_ties():
    # Test that ties are broken using the decision function,
    # not defaulting to the smallest label
    X = np.array([[1, 2], [2, 1], [-2, 1], [-2, -1]])
    y = np.array([2, 0, 1, 2])
    multi_clf = OneVsOneClassifier(Perceptron(shuffle=False, max_iter=4,
                                              tol=None))
    ovo_prediction = multi_clf.fit(X, y).predict(X)
    ovo_decision = multi_clf.decision_function(X)

    # Classifiers are in order 0-1, 0-2, 1-2
    # Use decision_function to compute the votes and the normalized
    # sum_of_confidences, which is used to disambiguate when there is a tie in
    # votes.
    votes = np.round(ovo_decision)
    normalized_confidences = ovo_decision - votes

    # For the first point, there is one vote per class
    assert_array_equal(votes[0, :], 1)
    # For the rest, there is no tie and the prediction is the argmax
    assert_array_equal(np.argmax(votes[1:], axis=1), ovo_prediction[1:])
    # For the tie, the prediction is the class with the highest score
    assert_equal(ovo_prediction[0], normalized_confidences[0].argmax()) 

def test_pairwise_attribute():
    clf_precomputed = svm.SVC(kernel='precomputed')
    clf_notprecomputed = svm.SVC()

    for MultiClassClassifier in [OneVsRestClassifier, OneVsOneClassifier]:
        ovr_false = MultiClassClassifier(clf_notprecomputed)
        assert_false(ovr_false._pairwise)

        ovr_true = MultiClassClassifier(clf_precomputed)
        assert_true(ovr_true._pairwise) 

def test_pairwise_indices():
    clf_precomputed = svm.SVC(kernel='precomputed')
    X, y = iris.data, iris.target

    ovr_false = OneVsOneClassifier(clf_precomputed)
    linear_kernel = np.dot(X, X.T)
    ovr_false.fit(linear_kernel, y)

    n_estimators = len(ovr_false.estimators_)
    precomputed_indices = ovr_false.pairwise_indices_

    for idx in precomputed_indices:
        assert_equal(idx.shape[0] * n_estimators / (n_estimators - 1),
                     linear_kernel.shape[0]) 

def test_ovo_one_class():
    # Test error for OvO with one class
    X = np.eye(4)
    y = np.array(['a'] * 4)

    ovo = OneVsOneClassifier(LinearSVC())
    assert_raise_message(ValueError, "when only one class", ovo.fit, X, y) 

def test_ovo_string_y():
    # Test that the OvO doesn't mess up the encoding of string labels
    X = np.eye(4)
    y = np.array(['a', 'b', 'c', 'd'])

    ovo = OneVsOneClassifier(LinearSVC())
    ovo.fit(X, y)
    assert_array_equal(y, ovo.predict(X)) 

def test_ovo_ties2():
    # test that ties can not only be won by the first two labels
    X = np.array([[1, 2], [2, 1], [-2, 1], [-2, -1]])
    y_ref = np.array([2, 0, 1, 2])

    # cycle through labels so that each label wins once
    for i in range(3):
        y = (y_ref + i) % 3
        multi_clf = OneVsOneClassifier(Perceptron(shuffle=False, max_iter=4,
                                                  tol=None))
        ovo_prediction = multi_clf.fit(X, y).predict(X)
        assert_equal(ovo_prediction[0], i % 3) 

def _one_vs_one(self,X,Y):
        self.cls = OneVsOneClassifier(KOMD(**self.get_params())).fit(X,Y)
        self.is_fitted = True
        return self 

def test_ovo_gridsearch():
    ovo = OneVsOneClassifier(LinearSVC(random_state=0))
    Cs = [0.1, 0.5, 0.8]
    cv = GridSearchCV(ovo, {'estimator__C': Cs})
    cv.fit(iris.data, iris.target)
    best_C = cv.best_estimator_.estimators_[0].C
    assert_true(best_C in Cs) 

def test_ovo_fit_predict():
    # A classifier which implements decision_function.
    ovo = OneVsOneClassifier(LinearSVC(random_state=0))
    ovo.fit(iris.data, iris.target).predict(iris.data)
    assert_equal(len(ovo.estimators_), n_classes * (n_classes - 1) / 2)

    # A classifier which implements predict_proba.
    ovo = OneVsOneClassifier(MultinomialNB())
    ovo.fit(iris.data, iris.target).predict(iris.data)
    assert_equal(len(ovo.estimators_), n_classes * (n_classes - 1) / 2) 

def test_ovo_fit_on_list():
    # Test that OneVsOne fitting works with a list of targets and yields the
    # same output as predict from an array
    ovo = OneVsOneClassifier(LinearSVC(random_state=0))
    prediction_from_array = ovo.fit(iris.data, iris.target).predict(iris.data)
    iris_data_list = [list(a) for a in iris.data]
    prediction_from_list = ovo.fit(iris_data_list,
                                   list(iris.target)).predict(iris_data_list)
    assert_array_equal(prediction_from_array, prediction_from_list) 

def test_ovo_exceptions():
    ovo = OneVsOneClassifier(LinearSVC(random_state=0))
    assert_raises(ValueError, ovo.predict, []) 

def test_pairwise_cross_val_score():
    clf_precomputed = svm.SVC(kernel='precomputed')
    clf_notprecomputed = svm.SVC(kernel='linear')

    X, y = iris.data, iris.target

    for MultiClassClassifier in [OneVsRestClassifier, OneVsOneClassifier]:
        ovr_false = MultiClassClassifier(clf_notprecomputed)
        ovr_true = MultiClassClassifier(clf_precomputed)

        linear_kernel = np.dot(X, X.T)
        score_precomputed = cross_val_score(ovr_true, linear_kernel, y)
        score_linear = cross_val_score(ovr_false, X, y)
        assert_array_equal(score_precomputed, score_linear) 

def onevsone_classifier(self, assign=False, **kwargs):
        """
        封装有监督学习分类器，使用OneVsOneClassifier进行多label的
        分类器二次封装，即：
             OneVsOneClassifier(self.clf, **kwargs)

        :param assign: 是否保存实例后的二次封装分类器对象，与其它构造器不同，
                       默认False，即默认不保存在类中替换原始分类器
        :param kwargs: 透传OneVsOneClassifier做为构造关键字参数
        :return: OneVsOneClassifier对象
        """
        onevsone = OneVsOneClassifier(self.clf, **kwargs)
        if assign:
            self.clf = onevsone
        return onevsone 

def test_objectmapper(self):
        df = pdml.ModelFrame([])
        self.assertIs(df.multiclass.OneVsRestClassifier, multiclass.OneVsRestClassifier)
        self.assertIs(df.multiclass.OneVsOneClassifier, multiclass.OneVsOneClassifier)
        self.assertIs(df.multiclass.OutputCodeClassifier, multiclass.OutputCodeClassifier) 

def __new__(cls, *args, multi_class="ovr", n_jobs=1, **kwargs):

        multiclass_meta = {"ovr": OneVsRestClassifier, "ovo": OneVsOneClassifier}[
            multi_class
        ]
        return multiclass_meta(
            _EqualOpportunityClassifier(*args, **kwargs), n_jobs=n_jobs
        ) 

def test_ovo_exceptions():
    ovo = OneVsOneClassifier(LinearSVC(random_state=0))
    assert_raises(ValueError, ovo.predict, []) 

def __init__(self, X, label_words): 
        # Encoding the labels (words to numbers) 
        self.le = preprocessing.LabelEncoder() 
 
        # Initialize One vs One Classifier using a linear kernel 
        self.clf = OneVsOneClassifier(LinearSVC(random_state=0)) 
 
        y = self._encodeLabels(label_words) 
        X = np.asarray(X) 
        self.clf.fit(X, y) 
 
    # Predict the output class for the input datapoint 

def main():
	"""
	Use a linear SVM for multi-class classification.
	
	One vs the rest : 77.61%
	Default         : 77.61%
	One vs one      : 85.07%
	"""
	
	seed = 123456789
	np.random.seed(seed)
	ntrain, ntest = 800, 200
	(tr_x, tr_y), (te_x, te_y) = load_mnist()
	x, y = np.vstack((tr_x, te_x)), np.hstack((tr_y, te_y))
	cv = MNISTCV(tr_y, te_y, ntrain, ntest, 1, seed)

	for tr, te in cv:
		clf = OneVsRestClassifier(LinearSVC(random_state=seed), -1)
		clf.fit(x[tr], y[tr])
		print clf.score(x[te], y[te])
		
		clf = LinearSVC(random_state=seed)
		clf.fit(x[tr], y[tr])
		print clf.score(x[te], y[te])
		
		clf = OneVsOneClassifier(LinearSVC(random_state=seed), -1)
		clf.fit(x[tr], y[tr])
		print clf.score(x[te], y[te]) 

def test_pairwise_cross_val_score():
    clf_precomputed = svm.SVC(kernel='precomputed')
    clf_notprecomputed = svm.SVC(kernel='linear')

    X, y = iris.data, iris.target

    for MultiClassClassifier in [OneVsRestClassifier, OneVsOneClassifier]:
        ovr_false = MultiClassClassifier(clf_notprecomputed)
        ovr_true = MultiClassClassifier(clf_precomputed)

        linear_kernel = np.dot(X, X.T)
        score_precomputed = cross_val_score(ovr_true, linear_kernel, y)
        score_linear = cross_val_score(ovr_false, X, y)
        assert_array_equal(score_precomputed, score_linear) 

def test_pairwise_attribute():
    clf_precomputed = svm.SVC(kernel='precomputed')
    clf_notprecomputed = svm.SVC()

    for MultiClassClassifier in [OneVsRestClassifier, OneVsOneClassifier]:
        ovr_false = MultiClassClassifier(clf_notprecomputed)
        assert not ovr_false._pairwise

        ovr_true = MultiClassClassifier(clf_precomputed)
        assert ovr_true._pairwise 

def test_pairwise_indices():
    clf_precomputed = svm.SVC(kernel='precomputed')
    X, y = iris.data, iris.target

    ovr_false = OneVsOneClassifier(clf_precomputed)
    linear_kernel = np.dot(X, X.T)
    ovr_false.fit(linear_kernel, y)

    n_estimators = len(ovr_false.estimators_)
    precomputed_indices = ovr_false.pairwise_indices_

    for idx in precomputed_indices:
        assert_equal(idx.shape[0] * n_estimators / (n_estimators - 1),
                     linear_kernel.shape[0]) 

def test_ovo_one_class():
    # Test error for OvO with one class
    X = np.eye(4)
    y = np.array(['a'] * 4)

    ovo = OneVsOneClassifier(LinearSVC())
    assert_raise_message(ValueError, "when only one class", ovo.fit, X, y) 

def __new__(cls, *args, multi_class="ovr", n_jobs=1, **kwargs):

        multiclass_meta = {"ovr": OneVsRestClassifier, "ovo": OneVsOneClassifier}[
            multi_class
        ]
        return multiclass_meta(
            _DemographicParityClassifer(*args, **kwargs), n_jobs=n_jobs
        ) 

def test_ovo_string_y():
    # Test that the OvO doesn't mess up the encoding of string labels
    X = np.eye(4)
    y = np.array(['a', 'b', 'c', 'd'])

    ovo = OneVsOneClassifier(LinearSVC())
    ovo.fit(X, y)
    assert_array_equal(y, ovo.predict(X)) 

def test_ovo_ties2():
    # test that ties can not only be won by the first two labels
    X = np.array([[1, 2], [2, 1], [-2, 1], [-2, -1]])
    y_ref = np.array([2, 0, 1, 2])

    # cycle through labels so that each label wins once
    for i in range(3):
        y = (y_ref + i) % 3
        multi_clf = OneVsOneClassifier(Perceptron(shuffle=False, max_iter=4,
                                                  tol=None))
        ovo_prediction = multi_clf.fit(X, y).predict(X)
        assert_equal(ovo_prediction[0], i % 3) 

def test_ovo_ties():
    # Test that ties are broken using the decision function,
    # not defaulting to the smallest label
    X = np.array([[1, 2], [2, 1], [-2, 1], [-2, -1]])
    y = np.array([2, 0, 1, 2])
    multi_clf = OneVsOneClassifier(Perceptron(shuffle=False, max_iter=4,
                                              tol=None))
    ovo_prediction = multi_clf.fit(X, y).predict(X)
    ovo_decision = multi_clf.decision_function(X)

    # Classifiers are in order 0-1, 0-2, 1-2
    # Use decision_function to compute the votes and the normalized
    # sum_of_confidences, which is used to disambiguate when there is a tie in
    # votes.
    votes = np.round(ovo_decision)
    normalized_confidences = ovo_decision - votes

    # For the first point, there is one vote per class
    assert_array_equal(votes[0, :], 1)
    # For the rest, there is no tie and the prediction is the argmax
    assert_array_equal(np.argmax(votes[1:], axis=1), ovo_prediction[1:])
    # For the tie, the prediction is the class with the highest score
    assert_equal(ovo_prediction[0], normalized_confidences[0].argmax())


# 0.23. warning about tol not having its correct default value. 

def test_ovo_gridsearch():
    ovo = OneVsOneClassifier(LinearSVC(random_state=0))
    Cs = [0.1, 0.5, 0.8]
    cv = GridSearchCV(ovo, {'estimator__C': Cs})
    cv.fit(iris.data, iris.target)
    best_C = cv.best_estimator_.estimators_[0].C
    assert best_C in Cs


# 0.23. warning about tol not having its correct default value. 

def test_ovo_fit_on_list():
    # Test that OneVsOne fitting works with a list of targets and yields the
    # same output as predict from an array
    ovo = OneVsOneClassifier(LinearSVC(random_state=0))
    prediction_from_array = ovo.fit(iris.data, iris.target).predict(iris.data)
    iris_data_list = [list(a) for a in iris.data]
    prediction_from_list = ovo.fit(iris_data_list,
                                   list(iris.target)).predict(iris_data_list)
    assert_array_equal(prediction_from_array, prediction_from_list)