GamBinaryTrainer.Options 类

定义

public sealed class GamBinaryTrainer.Options : Microsoft.ML.Trainers.FastTree.GamTrainerBase<Microsoft.ML.Trainers.FastTree.GamBinaryTrainer.Options,Microsoft.ML.Data.BinaryPredictionTransformer<Microsoft.ML.Calibrators.CalibratedModelParametersBase<Microsoft.ML.Trainers.FastTree.GamBinaryModelParameters,Microsoft.ML.Calibrators.PlattCalibrator>>,Microsoft.ML.Calibrators.CalibratedModelParametersBase<Microsoft.ML.Trainers.FastTree.GamBinaryModelParameters,Microsoft.ML.Calibrators.PlattCalibrator>>.OptionsBase
type GamBinaryTrainer.Options = class
    inherit GamTrainerBase<GamBinaryTrainer.Options, BinaryPredictionTransformer<CalibratedModelParametersBase<GamBinaryModelParameters, PlattCalibrator>>, CalibratedModelParametersBase<GamBinaryModelParameters, PlattCalibrator>>.OptionsBase
Public NotInheritable Class GamBinaryTrainer.Options
Inherits GamTrainerBase(Of GamBinaryTrainer.Options, BinaryPredictionTransformer(Of CalibratedModelParametersBase(Of GamBinaryModelParameters, PlattCalibrator)), CalibratedModelParametersBase(Of GamBinaryModelParameters, PlattCalibrator)).OptionsBase
继承

构造函数

GamBinaryTrainer.Options()

GamBinaryTrainer游戏 (选项) 中使用的选项

字段

DiskTranspose

无论是利用磁盘还是数据的本机转置设施 (在执行转置时) 。

(继承自 GamTrainerBase<TOptions,TTransformer,TPredictor>.OptionsBase)
EnablePruning

启用训练后树修剪以避免过度拟合。 它需要验证集。

(继承自 GamTrainerBase<TOptions,TTransformer,TPredictor>.OptionsBase)
EntropyCoefficient

entropy (正则化) 系数介于 0 和 1 之间。

(继承自 GamTrainerBase<TOptions,TTransformer,TPredictor>.OptionsBase)
ExampleWeightColumnName

要用于示例权重的列。

(继承自 TrainerInputBaseWithWeight)
FeatureColumnName

要用于功能的列。

(继承自 TrainerInputBase)
FeatureFlocks

是否在数据集准备期间收集特征,以加快训练速度。

(继承自 GamTrainerBase<TOptions,TTransformer,TPredictor>.OptionsBase)
GainConfidenceLevel

树拟合获得置信度要求。 仅当其可能性与随机选择增益高于此值时,才考虑增益。

(继承自 GamTrainerBase<TOptions,TTransformer,TPredictor>.OptionsBase)
GetDerivativesSampleRate

对 GetDerivatives 函数中的每个查询 1(以 k 次为单位)采样。

(继承自 GamTrainerBase<TOptions,TTransformer,TPredictor>.OptionsBase)
LabelColumnName

要用于标签的列。

(继承自 TrainerInputBaseWithLabel)
LearningRate

学习速率。

(继承自 GamTrainerBase<TOptions,TTransformer,TPredictor>.OptionsBase)
MaximumBinCountPerFeature

每个功能 (箱) 的最大非重复值数。

(继承自 GamTrainerBase<TOptions,TTransformer,TPredictor>.OptionsBase)
MaximumTreeOutput

单个树输出的绝对值的上限。

(继承自 GamTrainerBase<TOptions,TTransformer,TPredictor>.OptionsBase)
MinimumExampleCountPerLeaf

形成新树叶所需的最小数据点数。

(继承自 GamTrainerBase<TOptions,TTransformer,TPredictor>.OptionsBase)
NumberOfIterations

训练数据的传递总数。

(继承自 GamTrainerBase<TOptions,TTransformer,TPredictor>.OptionsBase)
NumberOfThreads

要使用的线程数。

(继承自 GamTrainerBase<TOptions,TTransformer,TPredictor>.OptionsBase)
Seed

随机数生成器的种子。

(继承自 GamTrainerBase<TOptions,TTransformer,TPredictor>.OptionsBase)
UnbalancedSets

是否使用针对不平衡训练数据优化的派生体。

适用于