共用方式為


DataDriftMetricThreshold 類別

注意

這是實驗性類別,可以隨時變更。 請參閱 https://aka.ms/azuremlexperimental 以取得詳細資訊。

資料漂移計量閾值

繼承
azure.ai.ml.entities._monitoring.thresholds.MetricThreshold
DataDriftMetricThreshold

建構函式

DataDriftMetricThreshold(*, data_type: ~typing.Literal[<MonitorFeatureType.CATEGORICAL: 'categorical'>, <MonitorFeatureType.NUMERICAL: 'numerical'>] = None, threshold: float = None, metric: str | None = None, numerical: ~azure.ai.ml.entities._monitoring.thresholds.NumericalDriftMetrics = None, categorical: ~azure.ai.ml.entities._monitoring.thresholds.CategoricalDriftMetrics = None)

參數

applicable_feature_type
Literal[ azure.ai.ml.constants.MonitorFeatureType.CATEGORICAL , azure.ai.ml.constants.MonitorFeatureType.MonitorFeatureType.NUMERICAL]
必要

計量閾值的功能類型

metric_name
Literal[ MonitorMetricName.JENSEN_SHANNON_DISTANCE , azure.ai.ml.constants.MonitorMetricName.NORMALIZED_WASSERSTEIN_DISTANCE , azure.ai.ml.constants.MonitorMetricName.POPULATION_STABILITY_INDEX , azure.ai.ml.constants.MonitorMetricName.TWO_SAMPLE_KOLMOGOROV_SMIRNOV_TEST , azure.ai.ml.constants.MonitorMetricName.PEARSONS_CHI_SQUARED_TEST]
必要

要計算的計量

threshold
float
必要

臨界值。 如果為 None,則會根據選取的計量來設定預設值。