Python sklearn.svm.SVC Examples

def multi_class_classification(data_X,data_Y):
    '''
    calculate multi-class classification and return related evaluation metrics
    '''

    svc = svm.SVC(C=1, kernel='linear')
    # X_train, X_test, y_train, y_test = train_test_split( data_X, data_Y, test_size=0.4, random_state=0) 
    clf = svc.fit(data_X, data_Y) #svm
    # array = svc.coef_
    # print array
    predicted = cross_val_predict(clf, data_X, data_Y, cv=2)
    print "accuracy",metrics.accuracy_score(data_Y, predicted)
    print "f1 score macro",metrics.f1_score(data_Y, predicted, average='macro') 
    print "f1 score micro",metrics.f1_score(data_Y, predicted, average='micro') 
    print "precision score",metrics.precision_score(data_Y, predicted, average='macro') 
    print "recall score",metrics.recall_score(data_Y, predicted, average='macro') 
    print "hamming_loss",metrics.hamming_loss(data_Y, predicted)
    print "classification_report", metrics.classification_report(data_Y, predicted)
    print "jaccard_similarity_score", metrics.jaccard_similarity_score(data_Y, predicted)
    # print "log_loss", metrics.log_loss(data_Y, predicted)
    print "zero_one_loss", metrics.zero_one_loss(data_Y, predicted)
    # print "AUC&ROC",metrics.roc_auc_score(data_Y, predicted)
    # print "matthews_corrcoef", metrics.matthews_corrcoef(data_Y, predicted) 

def _build_model(self,model_name,params=None):
        if params==None:
            if model_name=='xgb':
                self.model=XGBClassifier(n_estimators=100,learning_rate=0.02)
            elif model_name=='svm':
                kernel_function=chi2_kernel if not (self.model_kernel=='linear' or self.model_kernel=='rbf') else self.model_kernel
                self.model=SVC(C=1,kernel=kernel_function,gamma=1,probability=True)
            elif model_name=='lr':
                self.model=LR(C=1,penalty='l1',tol=1e-6)
        else:
            if model_name=='xgb':
                self.model=XGBClassifier(n_estimators=1000,learning_rate=0.02,**params)
            elif model_name=='svm':
                self.model=SVC(C=1,kernel=kernel_function,gamma=1,probability=True)
            elif model_name=='lr':
                self.model=LR(C=1,penalty='l1',tol=1e-6)

        log.l.info('=======> built the model {} done'.format(self.model_name)) 

def create_pandas_only_svm_classifier(X, y, probability=True):
    class PandasOnlyEstimator(TransformerMixin):
        def fit(self, X, y=None, **fitparams):
            return self

        def transform(self, X, **transformparams):
            dataset_is_df = isinstance(X, pd.DataFrame)
            if not dataset_is_df:
                raise Exception("Dataset must be a pandas dataframe!")
            return X

    pandas_only = PandasOnlyEstimator()

    clf = svm.SVC(gamma=0.001, C=100.0, probability=probability, random_state=777)
    pipeline = Pipeline([("pandas_only", pandas_only), ("clf", clf)])
    return pipeline.fit(X, y) 

def __init__(self, 
        learner=SVC(C=1000), 
        multiclass_strategy='ova', 
        verbose=False,
        max_iter=1000, 
        learning_rate=0.01, 
        tolerance=1e-7, 
        callbacks=[], 
        scheduler=None ):

        super().__init__(
            learner=learner, 
            multiclass_strategy=multiclass_strategy, 
            max_iter=max_iter, 
            verbose=verbose, 
            tolerance=tolerance,
            callbacks=callbacks,
            scheduler=scheduler, 
            direction='min', 
            learning_rate=learning_rate, 
        )
        self.func_form = summation 

def compute_accuracy_svc(
    data_train,
    labels_train,
    data_test,
    labels_test,
    param_grid=None,
    verbose=0,
    max_iter=-1,
):
    if param_grid is None:
        param_grid = [
            {"C": [1, 10, 100, 1000], "kernel": ["linear"]},
            {"C": [1, 10, 100, 1000], "gamma": [0.001, 0.0001], "kernel": ["rbf"]},
        ]
    svc = SVC(max_iter=max_iter)
    clf = GridSearchCV(svc, param_grid, verbose=verbose, cv=3)
    return compute_accuracy_classifier(
        clf, data_train, labels_train, data_test, labels_test
    ) 

def __init__(self, *args, **kwargs):
        super(MaximumLossReductionMaximalConfidence, self).__init__(*args, **kwargs)

        # self.n_labels = len(self.dataset.get_labeled_entries()[0][1])
        self.n_labels = len(self.dataset.get_labeled_entries()[1][0])

        random_state = kwargs.pop('random_state', None)
        self.random_state_ = seed_random_state(random_state)

        self.logreg_param = kwargs.pop('logreg_param',
                                       {'multi_class': 'multinomial',
                                        'solver': 'newton-cg',
                                        'random_state': random_state})
        self.logistic_regression_ = LogisticRegression(**self.logreg_param)

        self.br_base = kwargs.pop('br_base',
              SklearnProbaAdapter(SVC(kernel='linear',
                                      probability=True,
                                      gamma="auto",
                                      random_state=random_state))) 

def test_svm(self):
        svc_clf = SVC(gamma="auto")
        svc_clf.fit(self.X_train, self.y_train)
        svm = SVM()
        svm.train(Dataset(self.X_train, self.y_train))

        assert_array_equal(
            svc_clf.predict(self.X_train), svm.predict(self.X_train))
        assert_array_equal(
            svc_clf.predict(self.X_test), svm.predict(self.X_test))
        self.assertEqual(
            svc_clf.score(self.X_train, self.y_train),
            svm.score(Dataset(self.X_train, self.y_train)))
        self.assertEqual(
            svc_clf.score(self.X_test, self.y_test),
            svm.score(Dataset(self.X_test, self.y_test))) 

def objective_function(x, s):
    # Start the clock to determine the cost of this function evaluation
    start_time = time.time()

    # Shuffle the data and split up the request subset of the training data
    s_max = y_train.shape[0]
    shuffle = np.random.permutation(np.arange(s_max))
    train_subset = X_train[shuffle[:s]]
    train_targets_subset = y_train[shuffle[:s]]

    # Train the SVM on the subset set
    C = np.exp(float(x[0]))
    gamma = np.exp(float(x[1]))
    clf = svm.SVC(gamma=gamma, C=C)
    clf.fit(train_subset, train_targets_subset)
    
    # Validate this hyperparameter configuration on the full validation data
    y = 1 - clf.score(X_val, y_val)

    c = time.time() - start_time

    return y, c

# Load the data 

def buildModel(dataset, method, parameters):
    """
    Build final model for predicting real testing data
    """
    features = dataset.columns[0:-1]

    if method == 'RNN':
        clf = performRNNlass(dataset[features], dataset['UpDown'])
        return clf

    elif method == 'RF':
        clf = RandomForestClassifier(n_estimators=1000, n_jobs=-1)

    elif method == 'KNN':
        clf = neighbors.KNeighborsClassifier()

    elif method == 'SVM':
        c = parameters[0]
        g =  parameters[1]
        clf = SVC(C=c, gamma=g)

    elif method == 'ADA':
        clf = AdaBoostClassifier()

    return clf.fit(dataset[features], dataset['UpDown']) 

def sk_svm_train(intr, labeltr, inte, labelte, kener):
    clf = svm.SVC(kernel=kener)
    # 开始训练
    clf.fit(intr, labeltr)
    #  绘图的标识
    figsign = kener
    #  训练精确度
    acc_train = clf.score(intr, labeltr)
    #  测试精确度
    acc_test = clf.score(inte, labelte)
    #  支持向量的个数
    vec_count = sum(clf.n_support_)
    #  支持向量
    vectors = clf.support_vectors_

    return acc_train, acc_test, vec_count, vectors, figsign


# 结果输出函数 

def _prepare_classifier(self, params, n_jobs=1):

        X_train, y_train = params

        tuned_parameters = [{
            'kernel': ['rbf'], 
            'gamma': [1e-4,1e-3,1e-2,1e-1,1e+0,1e+1,1e+2,1e+3,1e+4],
            'C': [1e+0,1e+1,1e+2,1e+3,1e+4,1e+5,1e+6,1e+7,1e+8,1e+9]
        }]

        clf=RandomizedSearchCV(svm.SVC(random_state=self.random_state), 
                               tuned_parameters[0], 
                               n_iter=self.n_randomized_search_iter, 
                               n_jobs=n_jobs, random_state=self.random_state)
        clf.fit(X_train, y_train)
              
        params=clf.best_params_
        clf=svm.SVC(kernel=params['kernel'], C=params['C'], 
            gamma=params['gamma'], probability=True, 
            random_state=self.random_state)
        clf.fit(X_train, y_train)

        return clf 

def SVM():
    '''data1——线性分类'''
    data1 = spio.loadmat('data1.mat')
    X = data1['X']
    y = data1['y']
    y = np.ravel(y)
    plot_data(X, y)

    model = svm.SVC(C=1.0, kernel='linear').fit(X, y)  # 指定核函数为线性核函数
    plot_decisionBoundary(X, y, model)  # 画决策边界
    '''data2——非线性分类'''
    data2 = spio.loadmat('data2.mat')
    X = data2['X']
    y = data2['y']
    y = np.ravel(y)
    plt = plot_data(X, y)
    plt.show()

    model = svm.SVC(gamma=100).fit(X, y)  # gamma为核函数的系数，值越大拟合的越好
    plot_decisionBoundary(X, y, model, class_='notLinear')  # 画决策边界


# 作图 

def run_svms():
    svm_training_data, svm_validation_data, svm_test_data \
        = mnist_loader.load_data()
    accuracies = []
    for size in SIZES:
        print "\n\nTraining SVM with data set size %s" % size
        clf = svm.SVC()
        clf.fit(svm_training_data[0][:size], svm_training_data[1][:size])
        predictions = [int(a) for a in clf.predict(svm_validation_data[0])]
        accuracy = sum(int(a == y) for a, y in 
                       zip(predictions, svm_validation_data[1])) / 100.0
        print "Accuracy was %s percent" % accuracy
        accuracies.append(accuracy)
    f = open("more_data_svm.json", "w")
    json.dump(accuracies, f)
    f.close() 

def __call__(self, estimator):
        fitted_estimator = estimator.fit(self.X_train, self.y_train)

        if isinstance(estimator, (LinearClassifierMixin, SVC, NuSVC,
                                  LightBaseClassifier)):
            y_pred = estimator.decision_function(self.X_test)
        elif isinstance(estimator, DecisionTreeClassifier):
            y_pred = estimator.predict_proba(self.X_test.astype(np.float32))
        elif isinstance(
                estimator,
                (ForestClassifier, XGBClassifier, LGBMClassifier)):
            y_pred = estimator.predict_proba(self.X_test)
        else:
            y_pred = estimator.predict(self.X_test)

        return self.X_test, y_pred, fitted_estimator 

def test_linear_kernel():
    estimator = svm.SVC(kernel="linear", random_state=1)

    estimator.fit([[1], [2]], [1, 2])

    assembler = assemblers.SklearnSVMModelAssembler(estimator)
    actual = assembler.assemble()

    def kernel_ast(sup_vec_value):
        return ast.BinNumExpr(
            ast.NumVal(sup_vec_value),
            ast.FeatureRef(0),
            ast.BinNumOpType.MUL)

    expected = _create_expected_single_output_ast(
        estimator.dual_coef_, estimator.intercept_,
        [kernel_ast(1.0), kernel_ast(2.0)])

    assert utils.cmp_exprs(actual, expected) 

def test_classification():
    # Check classification for various parameter settings.
    rng = check_random_state(0)
    X_train, X_test, y_train, y_test = train_test_split(iris.data,
                                                        iris.target,
                                                        random_state=rng)
    grid = ParameterGrid({"max_samples": [0.5, 1.0],
                          "max_features": [1, 2, 4],
                          "bootstrap": [True, False],
                          "bootstrap_features": [True, False]})

    for base_estimator in [None,
                           DummyClassifier(),
                           Perceptron(tol=1e-3),
                           DecisionTreeClassifier(),
                           KNeighborsClassifier(),
                           SVC(gamma="scale")]:
        for params in grid:
            BaggingClassifier(base_estimator=base_estimator,
                              random_state=rng,
                              **params).fit(X_train, y_train).predict(X_test) 

def define_classification_model():
    """ Select and define the model you will use for the classifier. 
    """
    if config['model_type'] == 'linearSVM': # linearSVM can be faster than SVM
        return LinearSVC(C=1)
    elif config['model_type'] == 'SVM': # non-linearSVM, we can use the kernel trick
        return SVC(C=1, kernel='rbf', gamma='scale')
    elif config['model_type'] == 'kNN': # k-nearest neighbour
        return KNeighborsClassifier(n_neighbors=1, metric='cosine')
    elif config['model_type'] == 'perceptron': # otpimizes log-loss, also known as cross-entropy with sgd
        return SGDClassifier(max_iter=600, verbose=0.5, loss='log', learning_rate='optimal')
    elif config['model_type'] == 'MLP': # otpimizes log-loss, also known as cross-entropy with sgd
        return MLPClassifier(hidden_layer_sizes=(20,), max_iter=600, verbose=10, 
               solver='sgd', learning_rate='constant', learning_rate_init=0.001) 

def __init__(self, l=500, K=100, base_classifer=svm.SVC()):
        self.l = l
        self.K = K
        self.m = 0
        self.ut = None
        self.phi = 1
        self.base_classifer = base_classifer
        self.ix = None
        self._ix = None
        return 

def proxy_a_distance(source_X, target_X, verbose=False):
    """
    Compute the Proxy-A-Distance of a source/target representation
    """
    nb_source = np.shape(source_X)[0]
    nb_target = np.shape(target_X)[0]

    if verbose:
        print('PAD on', (nb_source, nb_target), 'examples')

    C_list = np.logspace(-5, 4, 10)

    half_source, half_target = int(nb_source/2), int(nb_target/2)
    train_X = np.vstack((source_X[0:half_source, :], target_X[0:half_target, :]))
    train_Y = np.hstack((np.zeros(half_source, dtype=int), np.ones(half_target, dtype=int)))

    test_X = np.vstack((source_X[half_source:, :], target_X[half_target:, :]))
    test_Y = np.hstack((np.zeros(nb_source - half_source, dtype=int), np.ones(nb_target - half_target, dtype=int)))

    best_risk = 1.0
    for C in C_list:
        clf = svm.SVC(C=C, kernel='linear', verbose=False)
        clf.fit(train_X, train_Y)

        train_risk = np.mean(clf.predict(train_X) != train_Y)
        test_risk = np.mean(clf.predict(test_X) != test_Y)

        if verbose:
            print('[ PAD C = %f ] train risk: %f  test risk: %f' % (C, train_risk, test_risk))

        if test_risk > .5:
            test_risk = 1. - test_risk

        best_risk = min(best_risk, test_risk)

    return 2 * (1. - 2 * best_risk) 

def create_sklearn_svm_classifier(X, y, probability=True):
    clf = svm.SVC(gamma=0.001, C=100.0, probability=probability, random_state=777)
    model = clf.fit(X, y)
    return model 

def train(self, depgraphs, modelfile):
        """
        :param depgraphs : list of DependencyGraph as the training data
        :type depgraphs : DependencyGraph
        :param modelfile : file name to save the trained model
        :type modelfile : str
        """

        try:
            input_file = tempfile.NamedTemporaryFile(
                prefix='transition_parse.train',
                dir=tempfile.gettempdir(),
                delete=False)

            if self._algorithm == self.ARC_STANDARD:
                self._create_training_examples_arc_std(depgraphs, input_file)
            else:
                self._create_training_examples_arc_eager(depgraphs, input_file)

            input_file.close()
            # Using the temporary file to train the libsvm classifier
            x_train, y_train = load_svmlight_file(input_file.name)
            # The parameter is set according to the paper:
            # Algorithms for Deterministic Incremental Dependency Parsing by Joakim Nivre
            # Todo : because of probability = True => very slow due to
            # cross-validation. Need to improve the speed here
            model = svm.SVC(
                kernel='poly',
                degree=2,
                coef0=0,
                gamma=0.2,
                C=0.5,
                verbose=True,
                probability=True)

            model.fit(x_train, y_train)
            # Save the model to file name (as pickle)
            pickle.dump(model, open(modelfile, 'wb'))
        finally:
            remove(input_file.name) 

def trainSVMwrapper(X,y):
    model = svm.SVC(kernel='linear', C=1)
    model.probability = True
    # model.class_weight = 'auto'
    model.fit(X, y)
    return model 

def pickC(self, X, y):
        Cs =      [1, 0.1, 5, 10, 50] # 1 goes first as it should be preferred
        scores = []
        n = X.shape[0]
        dev_index = max([int(n*0.8), 1+y.index(1)])
        max_score = 0.0
        self.C = Cs[0]
        print("Warning, not picking C from validation")
        return
        for i, C in enumerate(Cs) :
            this_model = svm.sparse.SVC(C=C, kernel='linear')
            this_model.probability = False
            this_model.class_weight = 'auto'
            
            this_model.fit(X[:dev_index,:],y[:dev_index])
            pred = this_model.predict(X)
            train_correct = 0.0
            dev_correct = 0.0
            for j, y_j in enumerate(y):
                if j < dev_index :
                    train_correct += int(y_j == pred[j])
                else :
                    dev_correct += int(y_j == pred[j])
            train_acc = train_correct/dev_index
            dev_acc = dev_correct/(n-dev_index)
            score = (0.1*train_acc + 0.9*dev_acc)
            print("\tfor C=%.2f;\n\t\t train_acc=%.4f, dev_acc=%.4f, score=%.4f" % (C, train_acc, dev_acc, score))
            if score > max_score :
                max_score = score
                self.C = C
            if score == 1.0 :
                break
        print("Selected C=%.2f"%self.C) 

def train(self, X, y):
        # print('train')
        # print(X[0])
        # print(type(X[0]))
        # print(numpy.shape(X))
        # print(y[0])
        self.pickC(X, y)
        #model = svm.sparse.SVC(kernel='linear', C=self.C)
        model = svm.SVC(kernel='linear', C=self.C)
        model.probability=True
        # model.class_weight = 'auto'
        model.fit(X,y)
        self.model = model 

def do_svm(train_set,train_labs,test_set,svmtype=0):
	if svmtype==2: probability=True
	else: probability = False
	clf = svm.SVC(kernel='linear',probability=probability)
	if svmtype==1: clf = svm.LinearSVC(loss='squared_hinge',penalty='l1',dual=False)
	clf.fit(train_set,train_labs)
	return clf.predict(test_set),clf 

def do_svm(train_set,train_labs,test_set,svmtype=0):
	if svmtype==2: probability=True
	else: probability = False
	clf = svm.SVC(kernel='linear',probability=probability)
	if svmtype==1: clf = svm.LinearSVC(loss='squared_hinge',penalty='l1',dual=False)
	clf.fit(train_set,train_labs)
	return clf.predict(test_set),clf 

def define_clfs_params(self):
        '''
        Defines all relevant parameters and classes for classfier objects.
        Edit these if you wish to change parameters.
        '''
        # These are the classifiers
        self.clfs = {
            'RF': RandomForestClassifier(n_estimators = 50, n_jobs = -1),
            'ET': ExtraTreesClassifier(n_estimators = 10, n_jobs = -1, criterion = 'entropy'),
            'AB': AdaBoostClassifier(DecisionTreeClassifier(max_depth = [1, 5, 10, 15]), algorithm = "SAMME", n_estimators = 200),
            'LR': LogisticRegression(penalty = 'l1', C = 1e5),
            'SVM': svm.SVC(kernel = 'linear', probability = True, random_state = 0),
            'GB': GradientBoostingClassifier(learning_rate = 0.05, subsample = 0.5, max_depth = 6, n_estimators = 10),
            'NB': GaussianNB(),
            'DT': DecisionTreeClassifier(),
            'SGD': SGDClassifier(loss = 'log', penalty = 'l2'),
            'KNN': KNeighborsClassifier(n_neighbors = 3)
            }
        # These are the parameters which will be run through
        self.params = {
             'RF':{'n_estimators': [1,10,100,1000], 'max_depth': [10, 15,20,30,40,50,60,70,100], 'max_features': ['sqrt','log2'],'min_samples_split': [2,5,10], 'random_state': [1]},
             'LR': {'penalty': ['l1','l2'], 'C': [0.00001,0.0001,0.001,0.01,0.1,1,10], 'random_state': [1]},
             'SGD': {'loss': ['log'], 'penalty': ['l2','l1','elasticnet'], 'random_state': [1]},
             'ET': {'n_estimators': [1,10,100,1000], 'criterion' : ['gini', 'entropy'], 'max_depth': [1,3,5,10,15], 'max_features': ['sqrt','log2'],'min_samples_split': [2,5,10], 'random_state': [1]},
             'AB': {'algorithm': ['SAMME', 'SAMME.R'], 'n_estimators': [1,10,100,1000], 'random_state': [1]},
             'GB': {'n_estimators': [1,10,100,1000], 'learning_rate' : [0.001,0.01,0.05,0.1,0.5],'subsample' : [0.1,0.5,1.0], 'max_depth': [1,3,5,10,20,50,100], 'random_state': [1]},
             'NB': {},
             'DT': {'criterion': ['gini', 'entropy'], 'max_depth': [1,2,15,20,30,40,50], 'max_features': ['sqrt','log2'],'min_samples_split': [2,5,10], 'random_state': [1]},
             'SVM' :{'C' :[0.00001,0.0001,0.001,0.01,0.1,1,10],'kernel':['linear'], 'random_state': [1]},
             'KNN' :{'n_neighbors': [1,5,10,25,50,100],'weights': ['uniform','distance'],'algorithm': ['auto','ball_tree','kd_tree']}
             } 

def __init__(self, model_type='classifier', n_ensemble=5):
        super(SVMQSAR, self).__init__()
        self.n_ensemble = n_ensemble
        self.model = []
        self.model_type = model_type
        if self.model_type == 'classifier':
            for i in range(n_ensemble):
                self.model.append(SVC())
        elif self.model_type == 'regressor':
            for i in range(n_ensemble):
                self.model.append(SVR())
        else:
            raise ValueError('invalid value for argument') 

def train_on_batch(self, domains, x, x_lens, y, y_lens, c=3000):
        """ Train svm on some data
        """
        self.model = svm.SVC(C=c, probability=True, verbose=2)
        examples = self.prepare_examples(x, x_lens, y, y_lens)
        labels = self.prepare_labels(domains)
        self.model.fit(examples, labels) 

def __init__(self, learner=SVC(C=1000), multiclass_strategy='ova', verbose=False,
				theta=0.0, max_iter=1000, learning_rate=0.01, callbacks=[]):
		super().__init__(learner=learner, generator=generator, multiclass_strategy=multiclass_strategy,	
			max_iter=max_iter, verbose=verbose, callbacks=callbacks)
		self.theta = theta
		self.func_form = summation
		print ('warning: MEMO needs refactoring and parameters chehck, please contact the author if you want to use MEMO')