Microsoftml Learners 개체

설명

다음 개체의 인스턴스는 모든 학습 함수에서 반환됩니다. 모두 BaseLearner 클래스에서 상속되고 공통 메서드를 구현합니다.

• get_algo_args는 학습 매개 변수를 반환하고,

• coef_는 계수를 검색합니다.

• summary_는 학습 정보를 반환합니다.

학습된 학습자에 따라 콘텐츠가 변경됩니다.

클래스 BaseLearner

microsoftml.modules.base_learner.BaseLearner(**kwargs)

모든 학습자에 대한 기본 클래스입니다.

coef_

모델 계수를 가져옵니다.

fit(formula: str, data: [revoscalepy.datasource.RxDataSource.RxDataSource,
    pandas.core.frame.DataFrame], ml_transforms: list = None,
    ml_transform_vars: list = None, row_selection: str = None,
    transforms: dict = None, transform_objects: dict = None,
    transform_function: str = None,
    transform_variables: list = None,
    transform_packages: list = None,
    transform_environment: dict = None, blocks_per_read: int = None,
    report_progress: int = None, verbose: int = 1,
    compute_context: revoscalepy.computecontext.RxComputeContext.RxComputeContext = None,
    **kargs)

모델을 맞춥니다.

get_algo_args()

알고리즘 인수를 가져옵니다.

predict(*args, **kwargs)

microsoftml.rx_predict().

summary_

모델 요약을 가져옵니다.

특정 학습자

FastTree 이진 또는 회귀 모델

microsoftml.FastTrees(method: ['binary', 'regression'] = 'binary',
    num_trees: int = 100, num_leaves: int = 20,
    learning_rate: float = 0.2, min_split: int = 10,
    example_fraction: float = 0.7, feature_fraction: float = 1,
    split_fraction: float = 1, num_bins: int = 255,
    first_use_penalty: float = 0, gain_conf_level: float = 0,
    unbalanced_sets: bool = False, train_threads: int = 8,
    random_seed: int = None,
    ensemble: microsoftml.modules.ensemble.EnsembleControl = None,
    **kargs)

학습 노드 가져오기

get_train_node(**all_args)

1클래스 svm

microsoftml.OneClassSvm(cache_size: float = 100,
    kernel: [<function linear_kernel at 0x0000007156EAC8C8>,
    <function polynomial_kernel at 0x0000007156EAC950>,
    <function rbf_kernel at 0x0000007156EAC7B8>,
    <function sigmoid_kernel at 0x0000007156EACA60>] = {'Name': 'RbfKernel',
    'Settings': {}}, epsilon: float = 0.001, nu: float = 0.1,
    shrink: bool = True, normalize: ['No', 'Warn', 'Auto',
    'Yes'] = 'Auto',
    ensemble: microsoftml.modules.ensemble.EnsembleControl = None,
    **kargs)

get_train_node(**all_args)

FastForest 이진 또는 회귀 모델

microsoftml.FastForest(method: ['binary', 'regression'] = 'binary',
    num_trees: int = 100, num_leaves: int = 20,
    min_split: int = 10, example_fraction: float = 0.7,
    feature_fraction: float = 0.7, split_fraction: float = 0.7,
    num_bins: int = 255, first_use_penalty: float = 0,
    gain_conf_level: float = 0, train_threads: int = 8,
    random_seed: int = None,
    ensemble: microsoftml.modules.ensemble.EnsembleControl = None,
    **kargs)

get_train_node(**all_args)

SDCA 이진 또는 회귀 모델

microsoftml.FastLinear(method: ['binary', 'regression'] = 'binary',
    loss_function: {'binary': [<function hinge_loss at 0x0000007156E8EA60>,
    <function log_loss at 0x0000007156E8E6A8>,
    <function smoothed_hinge_loss at 0x0000007156E8EAE8>],
    'regression': [<function squared_loss at 0x0000007156E8E950>]} = None,
    l2_weight: float = None, l1_weight: float = None,
    train_threads: int = None, convergence_tolerance: float = 0.1,
    max_iterations: int = None, shuffle: bool = True,
    check_frequency: int = None, normalize: ['No', 'Warn', 'Auto',
    'Yes'] = 'Auto',
    ensemble: microsoftml.modules.ensemble.EnsembleControl = None,
    **kargs)

get_train_node(**all_args)

로지스틱 회귀

microsoftml.LogisticRegression(method: ['binary',
    'multiClass'] = 'binary', l2_weight: float = 1,
    l1_weight: float = 1, opt_tol: float = 1e-07,
    memory_size: int = 20, init_wts_diameter: float = 0,
    max_iterations: int = 2147483647,
    show_training_stats: bool = False, sgd_init_tol: float = 0,
    train_threads: int = None, dense_optimizer: bool = False,
    normalize: ['No', 'Warn', 'Auto', 'Yes'] = 'Auto',
    ensemble: microsoftml.modules.ensemble.EnsembleControl = None,
    **kargs)

신경망

microsoftml.NeuralNetwork(method: ['binary', 'multiClass',
    'regression'] = 'binary', num_hidden_nodes: int = 100,
    num_iterations: int = 100, optimizer: ['adadelta_optimizer',
    'sgd_optimizer'] = {'Name': 'SgdOptimizer', 'Settings': {}},
    net_definition: str = None, init_wts_diameter: float = 0.1,
    max_norm: float = 0, acceleration: ['avx_math', 'clr_math',
    'gpu_math', 'mkl_math', 'sse_math'] = {'Name': 'AvxMath',
    'Settings': {}}, mini_batch_size: int = 1, normalize: ['No',
    'Warn', 'Auto', 'Yes'] = 'Auto',
    ensemble: microsoftml.modules.ensemble.EnsembleControl = None,
    **kargs)

get_train_node(**all_args)

모델 aic 가져오기

aic(k=2)

모델 계수 가져오기

coef_

잔차 편차 가져오기

deviance_

알고리즘 인수 가져오기

get_algo_args()

학습 노드 가져오기

get_train_node(**all_args)

관련 콘텐츠

rx_fast_forest, rx_fast_trees, rx_fast_linear, rx_logistic_regression, rx_neural_network, rx_oneclass_svmrx_predict