OlsModelParameters 类

定义

OlsTrainer模型参数。

public sealed class OlsModelParameters : Microsoft.ML.Trainers.RegressionModelParameters
type OlsModelParameters = class
    inherit RegressionModelParameters
Public NotInheritable Class OlsModelParameters
Inherits RegressionModelParameters
继承

字段

RSquared

确定系数。

RSquaredAdjusted

调整后的确定系数。 仅当模型中的示例多于参数加一个时,才能生成经过调整的 R 平方。 如果不满足此条件,此值将为 NaN

属性

Bias

预测器的偏差项。

(继承自 LinearModelParameters)
HasStatistics

模型是否具有每个参数的统计信息。 这是 false iff StandardErrorsTValuesPValues 均为 null。 模型可能没有每个参数的统计信息,因为模型中的示例数不多于参数,或者因为它们在训练中通过设置为 CalculateStatistics false 被显式禁止显示。

PValues

对应于每个模型标准错误的 p 值。 当且仅当 为 falseHasStatistics,此为 null

StandardErrors

每个模型参数的标准误差,其中第一个对应于偏差,所有后续参数依次对应于每个权重。 当且仅当 为 falseHasStatistics,此为 null

TValues

对应于每个模型标准错误的 t-Statistic 值。 当且仅当 为 falseHasStatistics,此为 null

Weights

预测器的特征权重系数。

(继承自 LinearModelParameters)

显式接口实现

ICalculateFeatureContribution.FeatureContributionCalculator

用于确定每个特征对示例分数的贡献 FeatureContributionCalculatingTransformer。 对于线性模型,给定特征的贡献等于特征值乘以相应权重的乘积。

(继承自 LinearModelParameters)
ICanSaveModel.Save(ModelSaveContext)

OlsTrainer模型参数。

(继承自 ModelParametersBase<TOutput>)

适用于