A templated (base) class for handling joint PDFs. More...

#include <JointPdf.h>

Inheritance diagram for QUESO::BaseJointPdf< V, M >:

Public Member Functions
Constructor/Destructor methods
	BaseJointPdf (const char *prefix, const VectorSet< V, M > &domainSet)
	Default constructor. More...

virtual	~BaseJointPdf ()
	Destructor. More...

Mathematical methods
virtual double	actualValue (const V &domainVector, const V domainDirection, V gradVector, M hessianMatrix, V hessianEffect) const =0
	Actual value of the PDF (scalar function). More...

virtual void	distributionMean (V &meanVector) const

virtual void	distributionVariance (M &covMatrix) const

virtual void	setNormalizationStyle (unsigned int value) const

void	setLogOfNormalizationFactor (double value) const
	Sets a logarithmic value to be used in the normalization factor (stored in the protected attribute m_normalizationStyle.) More...

virtual double	computeLogOfNormalizationFactor (unsigned int numSamples, bool m_logOfNormalizationFactor) const =0
	Computes the logarithm of the normalization factor. See template specialization. More...

virtual void	print (std::ostream &os) const
	Print method. Non-pure for backwards compatibility. More...

Public Member Functions inherited from QUESO::BaseScalarFunction< V, M >
void	setFiniteDifferenceStepSize (double fdStepSize)
	Sets the step size for finite differencing gradients. More...

void	setFiniteDifferenceStepSize (unsigned int i, double fdStepSize)

	BaseScalarFunction (const char *prefix, const VectorSet< V, M > &domainSet)
	Default constructor. More...

virtual	~BaseScalarFunction ()
	Destructor. More...

const VectorSet< V, M > &	domainSet () const
	Access to the protected attribute `m_domainSet:` domain set of the scalar function. More...

virtual double	lnValue (const V &domainVector, const V domainDirection, V gradVector, M hessianMatrix, V hessianEffect) const
	Logarithm of the value of the scalar function. Deprecated. More...

virtual double	lnValue (const V &domainVector) const
	Returns the logarithm of the function at `domainVector`. More...

virtual double	lnValue (const V &domainVector, V &gradVector) const
	Returns the logarithm of the function and its gradient at `domainVector`. More...

virtual double	lnValue (const V &domainVector, V &gradVector, const V &domainDirection, V &hessianEffect) const

Protected Member Functions
double	commonComputeLogOfNormalizationFactor (unsigned int numSamples, bool updateFactorInternally) const
	Common method (to the derived classes) to compute the logarithm of the normalization factor. More...

Protected Attributes
unsigned int	m_normalizationStyle

double	m_logOfNormalizationFactor

Protected Attributes inherited from QUESO::BaseScalarFunction< V, M >
const BaseEnvironment &	m_env

std::string	m_prefix

const VectorSet< V, M > &	m_domainSet
	Domain set of the scalar function. More...

Detailed Description

template<class V = GslVector, class M = GslMatrix>
class QUESO::BaseJointPdf< V, M >

A templated (base) class for handling joint PDFs.

This class allows the mathematical definition of a Joint PDF, which is a scalar function such as * \( \pi: B \subset R^n \rightarrow R \); ie a function of one or more variables that has always one-dimensional range. QUESO currently supports basic PDFs such as uniform and Gaussian and also more complex PDFs, such as the ones coming from a Bayesian analysis. They are implemented in the derived classes UniformJointPdf, GaussianJointPdf, and BayesianJointPdf, respectively. The posterior PDF may be represented within QUESO by GenericJointPdf.

Definition at line 55 of file JointPdf.h.

Constructor & Destructor Documentation

template<class V, class M>

QUESO::BaseJointPdf< V, M >::BaseJointPdf	(	const char *	prefix,
		const VectorSet< V, M > &	domainSet
	)

Default constructor.

Instantiates an object of the class, i.e. a scalar function, given a prefix and its domain.

Definition at line 33 of file JointPdf.C.

References QUESO::BaseEnvironment::displayVerbosity(), QUESO::BaseScalarFunction< V, M >::m_env, QUESO::BaseScalarFunction< V, M >::m_prefix, and QUESO::BaseEnvironment::subDisplayFile().

   :
   BaseScalarFunction<V,M>(((std::string)(prefix)+"pd_").c_str(), domainSet),
   m_normalizationStyle(0),
   m_logOfNormalizationFactor(0.)
 {
   if ((m_env.subDisplayFile()) && (m_env.displayVerbosity() >= 54)) {
     *m_env.subDisplayFile() << "Entering BaseJointPdf<V,M>::constructor() [3]"
                             << ": prefix = " << m_prefix
                             << std::endl;
   }
 
   if ((m_env.subDisplayFile()) && (m_env.displayVerbosity() >= 54)) {
     *m_env.subDisplayFile() << "Leaving BaseJointPdf<V,M>::constructor() [3]"
                             << ": prefix = " << m_prefix
                             << std::endl;
   }
 }

template<class V , class M >

QUESO::BaseJointPdf< V, M >::~BaseJointPdf ( )

virtual

Destructor.

Definition at line 55 of file JointPdf.C.

56 {

57 }

Member Function Documentation

template<class V = GslVector, class M = GslMatrix>

virtual double QUESO::BaseJointPdf< V, M >::actualValue	(	const V &	domainVector,
		const V *	domainDirection,
		V *	gradVector,
		M *	hessianMatrix,
		V *	hessianEffect
	)		const

pure virtual

Actual value of the PDF (scalar function).

Implements QUESO::BaseScalarFunction< V, M >.

Implemented in QUESO::InvLogitGaussianJointPdf< V, M >, QUESO::GammaJointPdf< V, M >, QUESO::ConcatenatedJointPdf< V, M >, QUESO::GaussianJointPdf< V, M >, QUESO::LogNormalJointPdf< V, M >, QUESO::BayesianJointPdf< V, M >, QUESO::PoweredJointPdf< V, M >, QUESO::InverseGammaJointPdf< V, M >, QUESO::GenericJointPdf< V, M >, QUESO::JeffreysJointPdf< V, M >, QUESO::BetaJointPdf< V, M >, QUESO::WignerJointPdf< V, M >, and QUESO::UniformJointPdf< V, M >.

template<class V , class M >

double QUESO::BaseJointPdf< V, M >::commonComputeLogOfNormalizationFactor	(	unsigned int	numSamples,
		bool	updateFactorInternally
	)		const

protected

Common method (to the derived classes) to compute the logarithm of the normalization factor.

The normalization factor is calculated by finding the max and min values of the domain set and then drawing numSamples samples from a uniform distribution varying from min to max. Such samples are averaged and the logarithmic value is assigned to protected attribute m_logOfNormalizationFactor if the parameter m_logOfNormalizationFactor is true.

Definition at line 115 of file JointPdf.C.

References QUESO::BoxSubset< V, M >::maxValues(), QUESO::BoxSubset< V, M >::minValues(), and QUESO::queso_isnan().

 {
   double value = 0.;
 
   double volume = m_domainSet.volume();
   if ((queso_isnan(volume)) ||
       (volume == -INFINITY         ) ||
       (volume ==  INFINITY         ) ||
       (volume <= 0.                )) {
     // Do nothing
   }
   else {
     const BoxSubset<V,M>* boxSubset = dynamic_cast<const BoxSubset<V,M>* >(&m_domainSet);
     if (boxSubset == NULL) {
       // Do nothing
     }
     else {
       V tmpVec(m_domainSet.vectorSpace().zeroVector());
       double sum = 0.;
       for (unsigned int i = 0; i < numSamples; ++i) {
         tmpVec.cwSetUniform(boxSubset->minValues(),boxSubset->maxValues());
         sum += this->actualValue(tmpVec,NULL,NULL,NULL,NULL);
       }
       double avgValue = sum/((double) numSamples);
       value = -( log(avgValue) + log(volume) );
       if (updateFactorInternally) {
         m_logOfNormalizationFactor = value;
       }
     }
   }
 
   return value;
 }

template<class V = GslVector, class M = GslMatrix>

virtual double QUESO::BaseJointPdf< V, M >::computeLogOfNormalizationFactor	(	unsigned int	numSamples,
		bool	m_logOfNormalizationFactor
	)		const

pure virtual

Computes the logarithm of the normalization factor. See template specialization.

template<class V, class M >

void QUESO::BaseJointPdf< V, M >::distributionMean ( V & meanVector ) const

virtual

Mean value of the underlying random variable. Not implemented in base class, but not pure virtual for backwards compatibility reasons.

Reimplemented in QUESO::InvLogitGaussianJointPdf< V, M >, QUESO::GammaJointPdf< V, M >, QUESO::GaussianJointPdf< V, M >, QUESO::ConcatenatedJointPdf< V, M >, QUESO::BayesianJointPdf< V, M >, QUESO::LogNormalJointPdf< V, M >, QUESO::JeffreysJointPdf< V, M >, QUESO::BetaJointPdf< V, M >, QUESO::InverseGammaJointPdf< V, M >, QUESO::WignerJointPdf< V, M >, and QUESO::UniformJointPdf< V, M >.

Definition at line 62 of file JointPdf.C.

 {
   queso_not_implemented();
 }

template<class V , class M>

void QUESO::BaseJointPdf< V, M >::distributionVariance ( M & covMatrix ) const

virtual

Covariance matrix of the underlying random variable. Not implemented in base class, but not pure virtual for backwards compatibility reasons.

Reimplemented in QUESO::InvLogitGaussianJointPdf< V, M >, QUESO::GammaJointPdf< V, M >, QUESO::GaussianJointPdf< V, M >, QUESO::ConcatenatedJointPdf< V, M >, QUESO::BayesianJointPdf< V, M >, QUESO::LogNormalJointPdf< V, M >, QUESO::JeffreysJointPdf< V, M >, QUESO::BetaJointPdf< V, M >, QUESO::InverseGammaJointPdf< V, M >, QUESO::WignerJointPdf< V, M >, and QUESO::UniformJointPdf< V, M >.

Definition at line 70 of file JointPdf.C.

 {
   queso_not_implemented();
 }

template<class V , class M >

void QUESO::BaseJointPdf< V, M >::print ( std::ostream & os ) const

virtual

Print method. Non-pure for backwards compatibility.

Reimplemented in QUESO::InvLogitGaussianJointPdf< V, M >, and QUESO::GaussianJointPdf< V, M >.

Definition at line 94 of file JointPdf.C.

 {
   // Print m_env?
   // Print mean?
   // Print var?
 
   os << "Start printing BaseJointPdf<V, M>" << std::endl;
   os << "m_prefix:" << std::endl;
   os << this->m_prefix << std::endl;
   os << "m_domainSet:" << std::endl;
   os << this->m_domainSet << std::endl;
   os << "m_normalizationStyle:" << std::endl;
   os << this->m_normalizationStyle << std::endl;
   os << "m_logOfNormalizationFactor:" << std::endl;
   os << this->m_logOfNormalizationFactor << std::endl;
   os << "End printing BaseJointPdf<V, M>" << std::endl;
 }

template<class V , class M >

void QUESO::BaseJointPdf< V, M >::setLogOfNormalizationFactor ( double value ) const

Sets a logarithmic value to be used in the normalization factor (stored in the protected attribute m_normalizationStyle.)

Definition at line 86 of file JointPdf.C.

 {
   m_logOfNormalizationFactor = value;
   return;
 }

template<class V , class M >

void QUESO::BaseJointPdf< V, M >::setNormalizationStyle ( unsigned int value ) const

virtual

Sets a value to be used in the normalization style. Default value is zero.

The value is stored in m_normalizationStyle.

If the normalization style is zero, one should compute analytical normalization constants in lnValue and actualValue. It doesn't appear to be used in many of of the computeLogOfNormalizationFactor methods in derived classes.

Reimplemented in QUESO::ConcatenatedJointPdf< V, M >, QUESO::BayesianJointPdf< V, M >, QUESO::BayesianJointPdf< QUESO::GslVector, QUESO::GslMatrix >, and QUESO::PoweredJointPdf< V, M >.

Definition at line 78 of file JointPdf.C.

 {
   m_normalizationStyle = value;
   return;
 }

Member Data Documentation

template<class V = GslVector, class M = GslMatrix>

double QUESO::BaseJointPdf< V, M >::m_logOfNormalizationFactor

mutableprotected

Definition at line 122 of file JointPdf.h.

template<class V = GslVector, class M = GslMatrix>

unsigned int QUESO::BaseJointPdf< V, M >::m_normalizationStyle

mutableprotected

Flag to decide which style of normalisation to use. The constructor sets this to zero in the initialisation list.

Definition at line 121 of file JointPdf.h.

The documentation for this class was generated from the following files:

src/stats/inc/JointPdf.h
src/stats/src/JointPdf.C

Public Member Functions

Protected Member Functions

Protected Attributes

Detailed Description

template<class V = GslVector, class M = GslMatrix> class QUESO::BaseJointPdf< V, M >

Constructor & Destructor Documentation

Member Function Documentation

Member Data Documentation

template<class V = GslVector, class M = GslMatrix>
class QUESO::BaseJointPdf< V, M >