cmp::prior::FromDistribution< DistType > Class Template Reference

Adapts a univariate probability distribution to act as a prior on a single coordinate. More...

#include <prior++.h>

Inheritance diagram for cmp::prior::FromDistribution< DistType >:

Collaboration diagram for cmp::prior::FromDistribution< DistType >:

Public Member Functions
	FromDistribution ()=delete
	FromDistribution (const FromDistribution &)=default
	FromDistribution (FromDistribution &&)=default
FromDistribution &	operator= (const FromDistribution &)=default
FromDistribution &	operator= (FromDistribution &&)=default
	FromDistribution (const DistType &dist, size_t paramIndex)
	Constructs a Prior from a Univariate Distribution.
double	eval (const Eigen::VectorXd &par) const override
double	evalGradient (const Eigen::VectorXd &par, const size_t &i) const override
double	evalHessian (const Eigen::VectorXd &par, const size_t &i, const size_t &j) const override
Public Member Functions inherited from cmp::prior::Prior
virtual	~Prior ()=default

Static Public Member Functions
static std::shared_ptr< Prior >	make (const DistType &dist, size_t paramIndex)

Private Attributes
DistType	dist_
	Underlying univariate distribution.
size_t	paramIndex_
	Parameter vector coordinate index.

Detailed Description

template<typename DistType>
class cmp::prior::FromDistribution< DistType >

Adapts a univariate probability distribution to act as a prior on a single coordinate.

Mathematical Formulation Applies a univariate distribution \( \mathcal{D} \) to a specific parameter dimension \( k \):

\[ \log p(\theta) = \log p_{\mathcal{D}}(\theta_k) \]

The gradient is non-zero only for coordinate \( k \):

\[ \frac{\partial \log p(\theta)}{\partial \theta_i} = \begin{cases} \frac{d \log p_{\mathcal{D}}(\theta_k)}{d \theta_k} & \text{if } i = k \\ 0 & \text{otherwise} \end{cases} \]

The Hessian is non-zero only for diagonal entry \( (k, k) \):

\[ \frac{\partial^2 \log p(\theta)}{\partial \theta_i \partial \theta_j} = \begin{cases} \frac{d^2 \log p_{\mathcal{D}}(\theta_k)}{d \theta_k^2} & \text{if } i = j = k \\ 0 & \text{otherwise} \end{cases} \]

Implementation Algorithm Accesses parameter coordinate \( \theta_k \) and delegates log-PDF and its first/second derivative calculations to the underlying dist_ object.

Constructor & Destructor Documentation

FromDistribution() [1/4]

template<typename DistType >

cmp::prior::FromDistribution< DistType >::FromDistribution ( )

delete

FromDistribution() [2/4]

template<typename DistType >

cmp::prior::FromDistribution< DistType >::FromDistribution ( const FromDistribution< DistType > &  )

default

FromDistribution() [3/4]

template<typename DistType >

cmp::prior::FromDistribution< DistType >::FromDistribution ( FromDistribution< DistType > &&  )

default

FromDistribution() [4/4]

template<typename DistType >

cmp::prior::FromDistribution< DistType >::FromDistribution ( const DistType &  dist, size_t  paramIndex  )

inline

Constructs a Prior from a Univariate Distribution.

Parameters

dist	The distribution instance (e.g., NormalDistribution)
paramIndex	The index in the Eigen::VectorXd this prior applies to

Member Function Documentation

eval()

template<typename DistType >

double cmp::prior::FromDistribution< DistType >::eval ( const Eigen::VectorXd &  par ) const

inlineoverridevirtual

Implements cmp::prior::Prior.

evalGradient()

template<typename DistType >

double cmp::prior::FromDistribution< DistType >::evalGradient ( const Eigen::VectorXd &  par, const size_t &  i  ) const

inlineoverridevirtual

Implements cmp::prior::Prior.

evalHessian()

template<typename DistType >

double cmp::prior::FromDistribution< DistType >::evalHessian ( const Eigen::VectorXd &  par, const size_t &  i, const size_t &  j  ) const

inlineoverridevirtual

Implements cmp::prior::Prior.

make()

template<typename DistType >

static std::shared_ptr< Prior > cmp::prior::FromDistribution< DistType >::make ( const DistType &  dist, size_t  paramIndex  )

inlinestatic

operator=() [1/2]

template<typename DistType >

FromDistribution & cmp::prior::FromDistribution< DistType >::operator= ( const FromDistribution< DistType > &  )

default

operator=() [2/2]

template<typename DistType >

FromDistribution & cmp::prior::FromDistribution< DistType >::operator= ( FromDistribution< DistType > &&  )

default

Member Data Documentation

dist_

template<typename DistType >

DistType cmp::prior::FromDistribution< DistType >::dist_

private

Underlying univariate distribution.

paramIndex_

template<typename DistType >

size_t cmp::prior::FromDistribution< DistType >::paramIndex_

private

Parameter vector coordinate index.

---

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

• include/prior++.h