संपादित करें

इसके माध्यम से साझा किया गया


Microsoftml Learners Objects

Description

An instance of the following objects is returned by every training function. They all inherit from the class BaseLearner and implement common methods.

  • get_algo_args returns the training parameters,

  • coef_retrieves the coefficients,

  • summary_ returns training information.

The content changes based on the trained learner.

Class BaseLearner

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

Base class for all the learners.

coef_

Get model coefficients.

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)

Fit the model.

get_algo_args()

Get algorithm arguments.

predict(*args, **kwargs)

Calls microsoftml.rx_predict().

summary_

Get model summary.

Specific Learners

FastTree binary or regression model

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)

Get train node

get_train_node(**all_args)

One-class 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 binary or regression model

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 binary or regression model

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)

Logistic regression

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)

Neural network

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)

Get model aic

aic(k=2)

Get model coefficients

coef_

Get residual deviance

deviance_

Get algorithm arguments

get_algo_args()

Get train node

get_train_node(**all_args)

rx_fast_forest, rx_fast_trees, rx_fast_linear, rx_logistic_regression, rx_neural_network, rx_oneclass_svm, rx_predict