likelihood Class Reference

Defines the combined likelihood class. More...

#include "likelihood/likelihood.h"

Inheritance diagram for likelihood:

Public Member Functions
	likelihood (std::vector< prior_base * > P, std::vector< transform_base * > T, std::vector< likelihood_base * > L, std::vector< double > &W)
	Likelihood class constructor.
	likelihood (std::vector< prior_base * > P, std::vector< likelihood_base * > L, std::vector< double > &W)
	Likelihood class constructor.
	~likelihood ()
	Likelihood class destructor.
virtual double	operator() (std::vector< double > &)
virtual std::vector< double >	gradient (std::vector< double > &)
	Gradient operator.
void	forward_transform (std::vector< double > &)
double	chi_squared (std::vector< double > &)
	Returns the chi squared evaluated at a vector in the parameter space.
bool	transform_state ()
void	use_finite_difference_gradients ()
	Set the likelihood to use finite differences in the gradient computation.
void	use_intrinsic_likelihood_gradients ()
	Set the likelihood to use the intrinsic likelihood gradients in the gradient computation.
std::vector< prior_base * >	priors ()
	Returns access to the prior vector.
std::vector< transform_base * >	transforms ()
	Returns access to the transform vector.
std::vector< likelihood_base * >	likelihoods ()
	Returns access to the base likelihood vector.
std::vector< double > &	weights ()
	Returns access to the weights vector.
void	set_mpi_communicator (MPI_Comm comm)
void	output_1d_slice (std::string fname, std::vector< double > p1, std::vector< double > p2, double xmin=-1, double xmax=2, size_t Nx=128)
	Output along a line.
void	output_2d_slice (std::string fname, std::vector< double > p1, std::vector< double > p2, std::vector< double > p3, double xmin=-1, double xmax=2, size_t Nx=128, double ymin=-1, double ymax=2, size_t Ny=128)
	Output a surface.

Protected Attributes
MPI_Comm	_comm

Private Member Functions
double	step_size (double u)

Private Attributes
std::vector< prior_base * >	_P
std::vector< transform_base * >	_T
std::vector< likelihood_base * >	_L
std::vector< double >	_W
std::vector< double >	_X
bool	_is_transform_provided
bool	_use_finite_difference_gradients

Detailed Description

Defines the combined likelihood class. This class combines different likelihoods into a single likelihood used by the samplers. It also contains the priors on all the parameters and the optional parameter transformations.

Member Function Documentation

void forward_transform

(

std::vector< double > &

x

)

Forward transformation function. It applies the parameter trasformations supplied to the constructor on the corresponding parameters

Here is the caller graph for this function:

double operator() ( std::vector< double > &  x )

virtual

Overloaded parenthesis operator that returns the log-likelihood of a vector in the parameter space at which the likelihood is to be calculated

Reimplemented in likelihood_power_tempered.

Here is the caller graph for this function:

void set_mpi_communicator

(

MPI_Comm

comm

)

Defines a set of processors provided to the model for parallel computation via an MPI communicator. Only facilates code parallelization if the model computation is parallelized via MPI.

Here is the caller graph for this function:

bool transform_state

(

)

Returns a boolean flag, "true" indicates that coordinate transformations are supplied to be carried on and "false" indicates no parameter transformation is performed

Here is the caller graph for this function:

Member Data Documentation

bool _is_transform_provided

private

Boolean variable, "true" indicates that coordinate transformations are supplied to be carried on and "false" indicates no parameter transformation is performed

std::vector<likelihood_base*> _L

private

A vector of pointers to likelihood objects, these are the individual likelihoods that combine to make the total likelihood

std::vector<prior_base*> _P

private

A vector of pointers to prior objects, these are the priors on each parameter in the parameter space

std::vector<transform_base*> _T

private

A vector of pointers to transformation objects, these are the coordinate transformations on each parameter in the parameter space

bool _use_finite_difference_gradients

private

Boolean switch to force full finite difference gradient computation at the combined likelihood level. If model evaluation is the limiting factor in gradient times and there are not major efficiencies to be found in the individual likelihoods, this should be set.

std::vector<double> _W

private

A vector of likelihood weights, these are the weights of individual likelihoods in the final combined likelihood

std::vector<double> _X

private

A vector in the parameter space, this is the transformed parameter vector if parameter transformations are supplied, otherwise it's the original vector of parameters

---

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

• likelihood/likelihood.h
• likelihood/likelihood.cpp