NonDBayesCalibration Class Reference

Base class for Bayesian inference: generates posterior distribution on model parameters given experimental data. More...

Inheritance diagram for NonDBayesCalibration:

List of all members.

Public Member Functions
	NonDBayesCalibration (ProblemDescDB &problem_db, Model &model)
	standard constructor
	~NonDBayesCalibration ()
	destructor
Protected Member Functions
void	derived_init_communicators (ParLevLIter pl_iter)
	derived class contributions to initializing the communicators associated with this Iterator instance
void	derived_set_communicators (ParLevLIter pl_iter)
	derived class contributions to setting the communicators associated with this Iterator instance
void	derived_free_communicators (ParLevLIter pl_iter)
	derived class contributions to freeing the communicators associated with this Iterator instance
const Model &	algorithm_space_model () const
void	initialize_model ()
	initialize emulator model and probability space transformations
Protected Attributes
short	emulatorType
	the emulator type: NO_EMULATOR, GP_EMULATOR, PCE_EMULATOR, or SC_EMULATOR
Model	mcmcModel
	Model instance employed in the likelihood function; provides response function values from Gaussian processes, stochastic expansions (PCE/SC), or direct access to simulations (no surrogate option)
Iterator	stochExpIterator
	NonDPolynomialChaos or NonDStochCollocation instance for defining a PCE/SC-based mcmcModel.
int	numSamples
	number of samples in the chain (e.g. number of MCMC samples)
int	chainCycles
	number of update cycles for MCMC chain (implemented by restarting of short chains)
int	randomSeed
	random seed for MCMC process
bool	standardizedSpace
	flag indicating use of a variable transformation to standardized probability space for the model or emulator
bool	adaptPosteriorRefine
	flag indicating usage of adaptive posterior refinement; currently makes sense for unstructured grids in GP and PCE least squares/CS
String	proposalCovarType
	approach for defining proposal covariance
RealVector	proposalCovarData
	data from user input of proposal covariance
String	proposalCovarFilename
	filename for user-specified proposal covariance
String	proposalCovarInputType
	approach for defining proposal covariance

---

Detailed Description

Base class for Bayesian inference: generates posterior distribution on model parameters given experimental data.

This class will eventually provide a general-purpose framework for Bayesian inference. In the short term, it only collects shared code between QUESO and GPMSA implementations.

---

Constructor & Destructor Documentation

NonDBayesCalibration	(	ProblemDescDB &	problem_db,
		Model &	model
	)

standard constructor

This constructor is called for a standard letter-envelope iterator instantiation. In this case, set_db_list_nodes has been called and probDescDB can be queried for settings from the method specification.

References Response::active_set(), NonDBayesCalibration::adaptPosteriorRefine, Iterator::algorithm_space_model(), Iterator::assign_rep(), Model::assign_rep(), NonD::cdfFlag, NonDBayesCalibration::chainCycles, Model::current_response(), NonDBayesCalibration::emulatorType, ProblemDescDB::get_bool(), ProblemDescDB::get_int(), ProblemDescDB::get_real(), ProblemDescDB::get_rv(), ProblemDescDB::get_string(), ProblemDescDB::get_usa(), ProblemDescDB::get_ushort(), Model::gradient_type(), Model::hessian_type(), NonD::initialize_random_variable_correlations(), NonD::initialize_random_variable_transformation(), NonD::initialize_random_variable_types(), Iterator::iteratedModel, Iterator::iterator_rep(), Iterator::maxEvalConcurrency, Iterator::maxIterations, NonDBayesCalibration::mcmcModel, NonDBayesCalibration::numSamples, Iterator::outputLevel, Iterator::probDescDB, NonDBayesCalibration::proposalCovarType, NonDBayesCalibration::randomSeed, ActiveSet::request_values(), NonD::requested_levels(), NonD::respLevelTarget, NonD::respLevelTargetReduce, NonDBayesCalibration::standardizedSpace, NonDBayesCalibration::stochExpIterator, NonD::transform_model(), and NonD::verify_correlation_support().

---

Member Function Documentation

const Model & algorithm_space_model

(

)

const [inline, protected, virtual]

default definition that gets redefined in selected derived Minimizers

Reimplemented from Analyzer.

References NonDBayesCalibration::mcmcModel.

---

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

• NonDBayesCalibration.hpp
• NonDBayesCalibration.cpp