Microsoft.ML.Transforms.TimeSeries 命名空间

包含时序数据转换组件的命名空间。

IidAnomalyDetectionBaseWrapper

这是计算所谓的 i.i.d 输入序列的 p 值和马丁加分的包装 Microsoft.ML.Transforms.TimeSeries.IidAnomalyDetectionBaseWrapper.IidAnomalyDetectionBase 器。 换句话说,它假定输入序列表示可能通过另一个进程计算的原始异常分数。

IidChangePointDetector

ITransformer 由拟合 a IidChangePointEstimator.

IidChangePointEstimator

根据自适应内核密度估计和马丁加尔,检测 独立相同分布 (i.i.d.) 时序上的信号更改。

IidSpikeDetector

ITransformer 由拟合 a IidSpikeEstimator.

IidSpikeEstimator

根据自适应内核密度估计,检测 独立相同分布 (i.i.d.) 时序上的信号峰值。

PredictionFunctionExtensions

包含时序数据转换组件的命名空间。

SrCnnAnomalyDetectionBase

包含时序数据转换组件的命名空间。

SrCnnAnomalyDetector

ITransformer 由拟合 a SrCnnAnomalyEstimator.

SrCnnAnomalyEstimator

使用谱残差 (SR) 算法检测时序中的异常

SsaAnomalyDetectionBaseWrapper

该包装器 Microsoft.ML.Transforms.TimeSeries.SsaAnomalyDetectionBaseWrapper.SsaAnomalyDetectionBase 基于时序的 Singular Spectrum 建模实现常规异常情况检测转换。 有关 SSA (SSA) 的奇异光谱分析的详细信息,请参阅 http://arxiv.org/pdf/1206.6910.pdf

SsaChangePointDetector

ITransformer 由拟合 a SsaChangePointEstimator.

SsaChangePointEstimator

使用单一光谱分析检测时序中的更改点。

SsaForecastingBaseWrapper

该包装器 Microsoft.ML.Transforms.TimeSeries.SsaForecastingBaseWrapper.SsaForecastingBase 基于时序的 Singular Spectrum 建模实现常规异常情况检测转换。 有关 SSA (SSA) 的奇异光谱分析的详细信息,请参阅 http://arxiv.org/pdf/1206.6910.pdf

SsaForecastingEstimator

使用单一光谱分析的预测。

SsaForecastingTransformer

ITransformer 由拟合 a SsaForecastingEstimator.

SsaSpikeDetector

ITransformer 由拟合 a SsaSpikeEstimator.

SsaSpikeEstimator

使用单一光谱分析检测时序峰值。

TimeSeriesPredictionEngine<TSrc,TDst>

运行以前训练的模型 (和上述转换管道) 内存中数据的类,一次一个示例。 这还可用于不以预测器结尾的已训练管道:在这种情况下,“预测”只是所有转换的结果。

结构

GrowthRatio

增长比率。 定义为 Growth^ (1/TimeSpan) 。

枚举

AnomalySide

异常情况检测的一侧。

ErrorFunction

包含时序数据转换组件的命名空间。

MartingaleType

马丁格尔的类型。

RankSelectionMethod

信号的排名选择方法。