StreamingFeaturizer 类

继承

builtins.object

StreamingFeaturizer

构造函数

StreamingFeaturizer(training_data: azureml.dataprep.api.dataflow.Dataflow, label_column: str, weight_column: Optional[str] = None, observer: Optional[azureml.automl.core._experiment_observer.ExperimentObserver] = None, featurization_config: Optional[azureml.automl.core.featurization.featurizationconfig.FeaturizationConfig] = None)

参数

training_data

label_column

weight_column

默认值: None

observer

默认值: None

featurization_config

默认值: None

方法

get_transformed_vector_column_names	获取转换后的工程特征名称。
get_transformers	获取流式处理特征化器的最终流式处理估算器。
learn_transformations	了解对训练数据的子样本的转换，然后根据整体进行拟合。

get_transformed_vector_column_names

获取转换后的工程特征名称。

get_transformed_vector_column_names() -> List[str]

返回

特征名称列表。

get_transformers

获取流式处理特征化器的最终流式处理估算器。

get_transformers() -> azureml.automl.runtime.featurization.streaming.streaming_estimator.StreamingEstimatorBase

返回

StreamingEstimatorBase

learn_transformations

了解对训练数据的子样本的转换，然后根据整体进行拟合。

learn_transformations() -> azureml.automl.runtime.featurization.streaming.streaming_featurization_transformer.StreamingFeaturizationTransformer

返回

StreamingFeaturizationTransformer

属性

FIT_ACTIVITY_NAME

FIT_ACTIVITY_NAME = 'StreamingFit'

MAX_NUM_BITS_FOR_ONE_HOT_HASH

MAX_NUM_BITS_FOR_ONE_HOT_HASH = 30

MAX_ROWS_TO_SUBSAMPLE

MAX_ROWS_TO_SUBSAMPLE = 100000

PARSE_FEATURE_NAMES_ACTIVITY_NAME

PARSE_FEATURE_NAMES_ACTIVITY_NAME = 'StreamingParseFeatureNames'