UniformNumericOption Class

Definition

Namespace:

Microsoft.ML.SearchSpace.Option

Assembly:

Microsoft.ML.SearchSpace.dll

Package:

Microsoft.ML.AutoML v0.21.1

abstract class for numeric option.

[System.Text.Json.Serialization.JsonConverter(typeof(Microsoft.ML.SearchSpace.Converter.NumericOptionConverter))]
public abstract class UniformNumericOption : Microsoft.ML.SearchSpace.Option.OptionBase

[<System.Text.Json.Serialization.JsonConverter(typeof(Microsoft.ML.SearchSpace.Converter.NumericOptionConverter))>]
type UniformNumericOption = class
    inherit OptionBase

Public MustInherit Class UniformNumericOption
Inherits OptionBase

Inheritance

Object

OptionBase

UniformNumericOption

Derived

Microsoft.ML.SearchSpace.Option.UniformDoubleOption

Microsoft.ML.SearchSpace.Option.UniformIntOption

Microsoft.ML.SearchSpace.Option.UniformSingleOption

Attributes

JsonConverterAttribute

Constructors

UniformNumericOption(Double, Double, Boolean)

Create a UniformNumericOption using min and max.

Properties

Default	Gets the default value which is mapping to feature space (if exists). (Inherited from OptionBase)
FeatureSpaceDim	the dimension of feature space, which is equal to the output length of SampleFromFeatureSpace(Double[]).
LogBase	Gets if this option is log base or not.
Max	Gets maximum value of this option.
Min	Gets minimum value of this option.
Step	Gets the step of this option. The Step is used to determine the number of grid this option should be divided into. In ChoiceOption, it's always the length of Choices. And in UniformNumericOption, it's always [null]. And in SearchSpace, it's a combination of all Step in its options.

Methods

MappingToFeatureSpace(Parameter)	mapping value to [0, 1) uniform distribution.
SampleFromFeatureSpace(Double[])	sample from [0,1) uniform distribution.