Class SparseGPOp

Provides outgoing messages for FunctionEvaluate(IFunction, Vector), given random arguments to the function.

Inheritance

Object

SparseGPOp

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

[FactorMethod(typeof(Factor), "FunctionEvaluate", new Type[]{typeof(IFunction), typeof(Vector)})]
[Quality(QualityBand.Preview)]
public static class SparseGPOp

Methods

AverageLogFactor(Double, IFunction, Vector)

Evidence message for VMP.

Declaration

public static double AverageLogFactor(double y, IFunction func, Vector x)

Parameters

Type	Name	Description
Double	y	Constant value for y.
IFunction	func	Constant value for func.
Vector	x	Constant value for x.

Returns

Type	Description
Double	Zero.

Remarks

The formula for the result is log(factor(y,func,x)). Adding up these values across all factors and variables gives the log-evidence estimate for VMP.

FuncAverageConditional(Gaussian, SparseGP, Vector, SparseGP)

EP message to func.

Declaration

public static SparseGP FuncAverageConditional(Gaussian y, SparseGP func, Vector x, SparseGP result)

Parameters

Type	Name	Description
Gaussian	y	Incoming message from y. Must be a proper distribution. If uniform, the result will be uniform.
SparseGP	func	Incoming message from func. Must be a proper distribution. If uniform, the result will be uniform.
Vector	x	Constant value for x.
SparseGP	result	Modified to contain the outgoing message.

Returns

Type	Description
SparseGP	result

Remarks

The outgoing message is a distribution matching the moments of func as the random arguments are varied. The formula is proj[p(func) sum_(y) p(y) factor(y,func,x)]/p(func).

Exceptions

Type	Condition
ImproperMessageException	y is not a proper distribution.
ImproperMessageException	func is not a proper distribution.

FuncAverageConditional(Double, SparseGP, Vector, SparseGP)

EP message to func.

Declaration

public static SparseGP FuncAverageConditional(double y, SparseGP func, Vector x, SparseGP result)

Parameters

Type	Name	Description
Double	y	Constant value for y.
SparseGP	func	Incoming message from func. Must be a proper distribution. If uniform, the result will be uniform.
Vector	x	Constant value for x.
SparseGP	result	Modified to contain the outgoing message.

Returns

Type	Description
SparseGP	result

Remarks

The outgoing message is the factor viewed as a function of func conditioned on the given values.

Exceptions

Type	Condition
ImproperMessageException	func is not a proper distribution.

LogAverageFactor(Double, IFunction, Vector)

Evidence message for EP.

Declaration

public static double LogAverageFactor(double y, IFunction func, Vector x)

Parameters

Type	Name	Description
Double	y	Constant value for y.
IFunction	func	Constant value for func.
Vector	x	Constant value for x.

Returns

Type	Description
Double	Logarithm of the factor's average value across the given argument distributions.

Remarks

The formula for the result is log(factor(y,func,x)).

LogAverageFactor(Double, SparseGP, Vector)

Evidence message for EP.

Declaration

public static double LogAverageFactor(double y, SparseGP func, Vector x)

Parameters

Type	Name	Description
Double	y	Constant value for y.
SparseGP	func	Incoming message from func.
Vector	x	Constant value for x.

Returns

Type	Description
Double	Logarithm of the factor's average value across the given argument distributions.

Remarks

The formula for the result is log(sum_(func) p(func) factor(y,func,x)).

LogEvidenceRatio(Gaussian)

Evidence message for EP.

Declaration

public static double LogEvidenceRatio(Gaussian y)

Parameters

Type	Name	Description
Gaussian	y	Incoming message from y.

Returns

Type	Description
Double	Logarithm of the factor's contribution the EP model evidence.

Remarks

The formula for the result is log(sum_(y) p(y) factor(y,func,x) / sum_y p(y) messageTo(y)). Adding up these values across all factors and variables gives the log-evidence estimate for EP.

LogEvidenceRatio(Double, IFunction, Vector)

Evidence message for EP.

Declaration

public static double LogEvidenceRatio(double y, IFunction func, Vector x)

Parameters

Type	Name	Description
Double	y	Constant value for y.
IFunction	func	Constant value for func.
Vector	x	Constant value for x.

Returns

Type	Description
Double	Logarithm of the factor's contribution the EP model evidence.

Remarks

The formula for the result is log(factor(y,func,x)). Adding up these values across all factors and variables gives the log-evidence estimate for EP.

LogEvidenceRatio(Double, SparseGP, Vector)

Evidence message for EP.

Declaration

public static double LogEvidenceRatio(double y, SparseGP func, Vector x)

Parameters

Type	Name	Description
Double	y	Constant value for y.
SparseGP	func	Incoming message from func.
Vector	x	Constant value for x.

Returns

Type	Description
Double	Logarithm of the factor's contribution the EP model evidence.

Remarks

The formula for the result is log(sum_(func) p(func) factor(y,func,x)). Adding up these values across all factors and variables gives the log-evidence estimate for EP.

YAverageConditional(SparseGP, Vector)

EP message to y.

Declaration

public static Gaussian YAverageConditional(SparseGP func, Vector x)

Parameters

Type	Name	Description
SparseGP	func	Incoming message from func. Must be a proper distribution. If uniform, the result will be uniform.
Vector	x	Constant value for x.

Returns

Type	Description
Gaussian	The outgoing EP message to the y argument.

Remarks

The outgoing message is a distribution matching the moments of y as the random arguments are varied. The formula is proj[p(y) sum_(func) p(func) factor(y,func,x)]/p(y).

Exceptions

Type	Condition
ImproperMessageException	func is not a proper distribution.