Class Damp

Damp factor methods

Inheritance

Object

Damp

Inherited Members

Object.Equals(Object)

Object.Equals(Object, Object)

Object.GetHashCode()

Object.GetType()

Object.MemberwiseClone()

Object.ReferenceEquals(Object, Object)

Object.ToString()

Namespace: Microsoft.ML.Probabilistic.Factors

Assembly: Microsoft.ML.Probabilistic.dll

Syntax

public static class Damp

Methods

Backward<T>(T, Double)

Copy a value and damp the message from the copy to the original.

Declaration

public static T Backward<T>(T value, double stepsize)

Parameters

Type	Name	Description
T	value
Double	stepsize	1.0 means no damping, 0.0 is infinite damping.

Returns

Type	Description
T

Type Parameters

Name	Description
T

Remarks

If you use this factor, be sure to multiply your convergence tolerance by stepsize. Equivalently, use the same convergence tolerance but measure the difference between iterations spaced apart by 1/stepsize.

Forward<T>(T, Double)

Copy a value and damp the message from the original to the copy.

Declaration

public static T Forward<T>(T value, double stepsize)

Parameters

Type	Name	Description
T	value
Double	stepsize	1.0 means no damping, 0.0 is infinite damping.

Returns

Type	Description
T

Type Parameters

Name	Description
T

Remarks

If you use this factor, be sure to multiply your convergence tolerance by stepsize. Equivalently, use the same convergence tolerance but measure the difference between iterations spaced apart by 1/stepsize.