Class DirichletSymmetricOp

Provides outgoing messages for DirichletSymmetric(Int32, Double), given random arguments to the function.

Inheritance

Object

DirichletSymmetricOp

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), "DirichletSymmetric", new Type[]{})]
[Quality(QualityBand.Preview)]
[Buffers(new string[]{"probMeanLog"})]
public static class DirichletSymmetricOp

Methods

AlphaAverageConditional(Gamma, Dirichlet, Gamma)

EP message to alpha.

Declaration

[NotSupported("Expectation Propagation does not currently support Dirichlet distributions with stochastic arguments")]
public static Gamma AlphaAverageConditional(Gamma alpha, Dirichlet prob, Gamma result)

Parameters

Type	Name	Description
Gamma	alpha	Incoming message from alpha. Must be a proper distribution. If uniform, the result will be uniform.
Dirichlet	prob	Incoming message from prob. Must be a proper distribution. If any element is uniform, the result will be uniform.
Gamma	result	Modified to contain the outgoing message.

Returns

Type	Description
Gamma	result

Remarks

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

Exceptions

Type	Condition
ImproperMessageException	alpha is not a proper distribution.
ImproperMessageException	prob is not a proper distribution.

AlphaAverageLogarithm(Dirichlet, ConjugateDirichlet, Vector)

VMP message to alpha.

Declaration

[Quality(QualityBand.Experimental)]
public static ConjugateDirichlet AlphaAverageLogarithm(Dirichlet prob, ConjugateDirichlet alpha, Vector probMeanLog)

Parameters

Type	Name	Description
Dirichlet	prob	Incoming message from prob. Must be a proper distribution. If any element is uniform, the result will be uniform.
ConjugateDirichlet	alpha	Incoming message from alpha. Must be a proper distribution. If uniform, the result will be uniform.
Vector	probMeanLog	Buffer probMeanLog.

Returns

Type	Description
ConjugateDirichlet	The outgoing VMP message to the alpha argument.

Remarks

The outgoing message is the exponential of the average log-factor value, where the average is over all arguments except alpha. The formula is exp(sum_(prob) p(prob) log(factor(prob,K,alpha))).

Exceptions

Type	Condition
ImproperMessageException	prob is not a proper distribution.
ImproperMessageException	alpha is not a proper distribution.

AlphaAverageLogarithm(Dirichlet, Gamma, Vector, Gamma)

VMP message to alpha.

Declaration

[Quality(QualityBand.Experimental)]
public static Gamma AlphaAverageLogarithm(Dirichlet prob, Gamma alpha, Vector probMeanLog, Gamma to_Alpha)

Parameters

Type	Name	Description
Dirichlet	prob	Incoming message from prob. Must be a proper distribution. If any element is uniform, the result will be uniform.
Gamma	alpha	Incoming message from alpha. Must be a proper distribution. If uniform, the result will be uniform.
Vector	probMeanLog	Buffer probMeanLog.
Gamma	to_Alpha	Previous outgoing message to Alpha.

Returns

Type	Description
Gamma	The outgoing VMP message to the alpha argument.

Remarks

The outgoing message is the exponential of the average log-factor value, where the average is over all arguments except alpha. The formula is exp(sum_(prob) p(prob) log(factor(prob,K,alpha))).

Exceptions

Type	Condition
ImproperMessageException	prob is not a proper distribution.
ImproperMessageException	alpha is not a proper distribution.

AlphaAverageLogarithm(Vector, ConjugateDirichlet, Vector)

VMP message to alpha.

Declaration

[Quality(QualityBand.Experimental)]
public static ConjugateDirichlet AlphaAverageLogarithm(Vector prob, ConjugateDirichlet alpha, Vector probMeanLog)

Parameters

Type	Name	Description
Vector	prob	Constant value for prob.
ConjugateDirichlet	alpha	Incoming message from alpha.
Vector	probMeanLog	Buffer probMeanLog.

Returns

Type	Description
ConjugateDirichlet	The outgoing VMP message to the alpha argument.

Remarks

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

AlphaAverageLogarithm(Vector, Gamma, Gamma)

VMP message to alpha.

Declaration

[Quality(QualityBand.Experimental)]
public static Gamma AlphaAverageLogarithm(Vector prob, Gamma alpha, Gamma to_Alpha)

Parameters

Type	Name	Description
Vector	prob	Constant value for prob.
Gamma	alpha	Incoming message from alpha. Must be a proper distribution. If uniform, the result will be uniform.
Gamma	to_Alpha	Previous outgoing message to Alpha.

Returns

Type	Description
Gamma	The outgoing VMP message to the alpha argument.

Remarks

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

Exceptions

Type	Condition
ImproperMessageException	alpha is not a proper distribution.

AverageLogFactor(Dirichlet, ConjugateDirichlet)

Evidence message for VMP.

Declaration

[Quality(QualityBand.Experimental)]
public static double AverageLogFactor(Dirichlet prob, ConjugateDirichlet alpha)

Parameters

Type	Name	Description
Dirichlet	prob	Incoming message from prob. Must be a proper distribution. If any element is uniform, the result will be uniform.
ConjugateDirichlet	alpha	Incoming message from alpha.

Returns

Type	Description
Double	Average of the factor's log-value across the given argument distributions.

Remarks

The formula for the result is sum_(prob,alpha) p(prob,alpha) log(factor(prob,K,alpha)). Adding up these values across all factors and variables gives the log-evidence estimate for VMP.

Exceptions

Type	Condition
ImproperMessageException	prob is not a proper distribution.

AverageLogFactor(Dirichlet, Gamma, Vector)

Evidence message for VMP.

Declaration

[Quality(QualityBand.Experimental)]
public static double AverageLogFactor(Dirichlet prob, Gamma alpha, Vector probMeanLog)

Parameters

Type	Name	Description
Dirichlet	prob	Incoming message from prob. Must be a proper distribution. If any element is uniform, the result will be uniform.
Gamma	alpha	Incoming message from alpha. Must be a proper distribution. If uniform, the result will be uniform.
Vector	probMeanLog	Buffer probMeanLog.

Returns

Type	Description
Double	Average of the factor's log-value across the given argument distributions.

Remarks

The formula for the result is sum_(prob,alpha) p(prob,alpha) log(factor(prob,K,alpha)). Adding up these values across all factors and variables gives the log-evidence estimate for VMP.

Exceptions

Type	Condition
ImproperMessageException	prob is not a proper distribution.
ImproperMessageException	alpha is not a proper distribution.

AverageLogFactor(Dirichlet, Double, Vector)

Evidence message for VMP.

Declaration

[Quality(QualityBand.Stable)]
public static double AverageLogFactor(Dirichlet prob, double alpha, Vector probMeanLog)

Parameters

Type	Name	Description
Dirichlet	prob	Incoming message from prob. Must be a proper distribution. If any element is uniform, the result will be uniform.
Double	alpha	Constant value for alpha.
Vector	probMeanLog	Buffer probMeanLog.

Returns

Type	Description
Double	Average of the factor's log-value across the given argument distributions.

Remarks

The formula for the result is sum_(prob) p(prob) log(factor(prob,K,alpha)). Adding up these values across all factors and variables gives the log-evidence estimate for VMP.

Exceptions

Type	Condition
ImproperMessageException	prob is not a proper distribution.

AverageLogFactor(Vector, ConjugateDirichlet)

Evidence message for VMP.

Declaration

[Quality(QualityBand.Experimental)]
public static double AverageLogFactor(Vector prob, ConjugateDirichlet alpha)

Parameters

Type	Name	Description
Vector	prob	Constant value for prob.
ConjugateDirichlet	alpha	Incoming message from alpha. Must be a proper distribution. If uniform, the result will be uniform.

Returns

Type	Description
Double	Average of the factor's log-value across the given argument distributions.

Remarks

The formula for the result is sum_(alpha) p(alpha) log(factor(prob,K,alpha)). Adding up these values across all factors and variables gives the log-evidence estimate for VMP.

Exceptions

Type	Condition
ImproperMessageException	alpha is not a proper distribution.

AverageLogFactor(Vector, Gamma)

Evidence message for VMP.

Declaration

[Quality(QualityBand.Experimental)]
public static double AverageLogFactor(Vector prob, Gamma alpha)

Parameters

Type	Name	Description
Vector	prob	Constant value for prob.
Gamma	alpha	Incoming message from alpha. Must be a proper distribution. If uniform, the result will be uniform.

Returns

Type	Description
Double	Average of the factor's log-value across the given argument distributions.

Remarks

The formula for the result is sum_(alpha) p(alpha) log(factor(prob,K,alpha)). Adding up these values across all factors and variables gives the log-evidence estimate for VMP.

Exceptions

Type	Condition
ImproperMessageException	alpha is not a proper distribution.

AverageLogFactor(IList<Double>, Double)

Evidence message for VMP.

Declaration

public static double AverageLogFactor(IList<double> prob, double alpha)

Parameters

Type	Name	Description
IList<Double>	prob	Incoming message from prob.
Double	alpha	Constant value for alpha.

Returns

Type	Description
Double	Average of the factor's log-value across the given argument distributions.

Remarks

The formula for the result is sum_(prob) p(prob) log(factor(prob,K,alpha)). Adding up these values across all factors and variables gives the log-evidence estimate for VMP.

LogAverageFactor(Dirichlet, Gamma)

Evidence message for EP.

Declaration

[NotSupported("Expectation Propagation does not currently support Dirichlet distributions with stochastic arguments")]
public static double LogAverageFactor(Dirichlet prob, Gamma alpha)

Parameters

Type	Name	Description
Dirichlet	prob	Incoming message from prob. Must be a proper distribution. If any element is uniform, the result will be uniform.
Gamma	alpha	Incoming message from alpha. Must be a proper distribution. If uniform, the result will be uniform.

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_(prob,alpha) p(prob,alpha) factor(prob,K,alpha)).

Exceptions

Type	Condition
ImproperMessageException	prob is not a proper distribution.
ImproperMessageException	alpha is not a proper distribution.

LogAverageFactor(Dirichlet, Double)

Evidence message for EP.

Declaration

public static double LogAverageFactor(Dirichlet prob, double alpha)

Parameters

Type	Name	Description
Dirichlet	prob	Incoming message from prob.
Double	alpha	Constant value for alpha.

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_(prob) p(prob) factor(prob,K,alpha)).

LogAverageFactor(IList<Double>, Double)

Evidence message for EP.

Declaration

public static double LogAverageFactor(IList<double> prob, double alpha)

Parameters

Type	Name	Description
IList<Double>	prob	Incoming message from prob.
Double	alpha	Constant value for alpha.

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_(prob) p(prob) factor(prob,K,alpha)).

LogEvidenceRatio(Dirichlet, Double)

Evidence message for EP.

Declaration

public static double LogEvidenceRatio(Dirichlet prob, double alpha)

Parameters

Type	Name	Description
Dirichlet	prob	Incoming message from prob.
Double	alpha	Constant value for alpha.

Returns

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

Remarks

The formula for the result is log(sum_(prob) p(prob) factor(prob,K,alpha) / sum_prob p(prob) messageTo(prob)). Adding up these values across all factors and variables gives the log-evidence estimate for EP.

LogEvidenceRatio(IList<Double>, Double)

Evidence message for EP.

Declaration

public static double LogEvidenceRatio(IList<double> prob, double alpha)

Parameters

Type	Name	Description
IList<Double>	prob	Incoming message from prob.
Double	alpha	Constant value for alpha.

Returns

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

Remarks

The formula for the result is log(sum_(prob) p(prob) factor(prob,K,alpha) / sum_prob p(prob) messageTo(prob)). Adding up these values across all factors and variables gives the log-evidence estimate for EP.

ProbAverageConditional(Gamma, Dirichlet)

EP message to prob.

Declaration

[NotSupported("Expectation Propagation does not currently support Dirichlet distributions with stochastic arguments")]
public static Dirichlet ProbAverageConditional(Gamma alpha, Dirichlet result)

Parameters

Type	Name	Description
Gamma	alpha	Incoming message from alpha. Must be a proper distribution. If uniform, the result will be uniform.
Dirichlet	result	Modified to contain the outgoing message.

Returns

Type	Description
Dirichlet	result

Remarks

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

Exceptions

Type	Condition
ImproperMessageException	alpha is not a proper distribution.

ProbAverageConditional(Double, Dirichlet)

EP message to prob.

Declaration

public static Dirichlet ProbAverageConditional(double alpha, Dirichlet result)

Parameters

Type	Name	Description
Double	alpha	Constant value for alpha.
Dirichlet	result	Modified to contain the outgoing message.

Returns

Type	Description
Dirichlet	result

Remarks

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

ProbAverageConditionalInit(Int32)

Declaration

public static Dirichlet ProbAverageConditionalInit(int K)

Parameters

Type	Name	Description
Int32	K	Constant value for K.

Returns

Type	Description
Dirichlet

Remarks

ProbAverageLogarithm(ConjugateDirichlet, Dirichlet)

VMP message to prob.

Declaration

[Quality(QualityBand.Experimental)]
public static Dirichlet ProbAverageLogarithm(ConjugateDirichlet alpha, Dirichlet result)

Parameters

Type	Name	Description
ConjugateDirichlet	alpha	Incoming message from alpha. Must be a proper distribution. If uniform, the result will be uniform.
Dirichlet	result	Modified to contain the outgoing message.

Returns

Type	Description
Dirichlet	result

Remarks

The outgoing message is the exponential of the average log-factor value, where the average is over all arguments except prob. The formula is exp(sum_(alpha) p(alpha) log(factor(prob,K,alpha))).

Exceptions

Type	Condition
ImproperMessageException	alpha is not a proper distribution.

ProbAverageLogarithm(Gamma, Dirichlet)

VMP message to prob.

Declaration

[Quality(QualityBand.Experimental)]
public static Dirichlet ProbAverageLogarithm(Gamma alpha, Dirichlet result)

Parameters

Type	Name	Description
Gamma	alpha	Incoming message from alpha. Must be a proper distribution. If uniform, the result will be uniform.
Dirichlet	result	Modified to contain the outgoing message.

Returns

Type	Description
Dirichlet	result

Remarks

The outgoing message is the exponential of the average log-factor value, where the average is over all arguments except prob. The formula is exp(sum_(alpha) p(alpha) log(factor(prob,K,alpha))).

Exceptions

Type	Condition
ImproperMessageException	alpha is not a proper distribution.

ProbAverageLogarithm(Double, Dirichlet)

VMP message to prob.

Declaration

[Quality(QualityBand.Stable)]
public static Dirichlet ProbAverageLogarithm(double alpha, Dirichlet result)

Parameters

Type	Name	Description
Double	alpha	Constant value for alpha.
Dirichlet	result	Modified to contain the outgoing message.

Returns

Type	Description
Dirichlet	result

Remarks

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

ProbAverageLogarithmInit(Int32)

Declaration

public static Dirichlet ProbAverageLogarithmInit(int K)

Parameters

Type	Name	Description
Int32	K	Constant value for K.

Returns

Type	Description
Dirichlet

Remarks

ProbMeanLog(Dirichlet, Vector)

Update the buffer probMeanLog.

Declaration

public static Vector ProbMeanLog(Dirichlet prob, Vector result)

Parameters

Type	Name	Description
Dirichlet	prob	Incoming message from prob.
Vector	result	Modified to contain the outgoing message.

Returns

Type	Description
Vector	result

Remarks

ProbMeanLog(Vector, Vector)

Update the buffer probMeanLog.

Declaration

public static Vector ProbMeanLog(Vector prob, Vector result)

Parameters

Type	Name	Description
Vector	prob	Constant value for prob.
Vector	result	Modified to contain the outgoing message.

Returns

Type	Description
Vector	result

Remarks

ProbMeanLogInit(Dirichlet)

Initialize the buffer probMeanLog.

Declaration

public static Vector ProbMeanLogInit(Dirichlet prob)

Parameters

Type	Name	Description
Dirichlet	prob	Incoming message from prob.

Returns

Type	Description
Vector	Initial value of buffer probMeanLog.

Remarks

ProbMeanLogInit(Vector)

Initialize the buffer probMeanLog.

Declaration

public static Vector ProbMeanLogInit(Vector prob)

Parameters

Type	Name	Description
Vector	prob	Constant value for prob.

Returns

Type	Description
Vector	Initial value of buffer probMeanLog.

Remarks