BIE::PosteriorProb Class Reference

Print out value of posterior probability for a prompted input state. More...

#include <PosteriorProb.h>

Inheritance diagram for BIE::PosteriorProb:

[legend]

Collaboration diagram for BIE::PosteriorProb:

[legend]

List of all members.

Public Types
enum	Control { parallel, serial }
	Parallelization type. More...
Public Member Functions
	PosteriorProb (int nmix, int ndim, BaseDataTree *d, Model *m, Integration *i, MixturePrior *mp, LikelihoodComputation *l)
	Constructor.
	~PosteriorProb ()
	Destructor.
void	Run ()
	Determine a new state.
void	Run (string)
	Determine a new state reading input from a file.
void	setLevels (int Ndecend)
	Set frontier.
void	SetControl (int)
	Choose parallelization control.
void	Image (string filename)
	Write out an image cut of the posterior.
void	Image (string filename, string input)
	Write out an image cut of the posterior.
void	SaveFrontier ()
	Save the current frontier for likelihood ratio.
void	NullFrontier ()
	Reset the prior frontier to null (no likelihood ratio).
Static Public Attributes
static Control	cntrl
	Current control.
static Converge	c
	Not relevant.
Protected Member Functions
void	MCMethod ()
	Execute the Markov Chain algorithm.
void	GetUserState ()
	Request state from user.
void	ReadState (string)
	Request state from file.
Utility conversions
vector< double >	mix_to_State (vector< double > &, vector< double > *&)
	Convert the weights and parameter vectors to a State.
void	State_to_mix (vector< double > &s, vector< double > &w, vector< double > *&p)
	Convert a State to weights and parameter vectors.
void	ReportState ()
	Log the current state to the simulation output stream.
void	PrintState ()
	Print the current state to the console.
Utility routines for Image methods
bool	GetImageDataUser (int &numx, int &numy, vector< double > &wght0, vector< vector< double > > &phi0, vector< char > &bwght, vector< vector< char > > &bphi, vector< vector< double > > &range)
bool	GetImageDataFile (string &input, int &numx, int &numy, vector< double > &wght0, vector< vector< double > > &phi0, vector< char > &bwght, vector< vector< char > > &bphi, vector< vector< double > > &range)
void	ShareImageData (int &numx, int &numy, vector< double > &wght0, vector< vector< double > > &phi0, vector< char > &bwght, vector< vector< char > > &bphi, vector< vector< double > > &range)
Protected Attributes
MixturePrior *	_mprior
	Pointer to the current prior distribution.
Friends
class	boost::serialization::access

---

Detailed Description

Print out value of posterior probability for a prompted input state.

The class is designed to be used interactively for exploring the sensitivity of the posterior probability to parameters. By setting "debugflags" to report bad cells, one can the investigate failures of the model to explain a data set.

---

Member Enumeration Documentation

enum BIE::PosteriorProb::Control

Parallelization type.

Enumerator:

parallel	root controls all chain updates.
serial	chains are run at each node.

---

Member Function Documentation

void BIE::PosteriorProb::SetControl

(

int

)

Choose parallelization control.

(parallel=0, serial=1)

1. The parallel likelihood classes should be used with parallel control only. The parallel likelihood classes LikelihoodComputationMPI and LikelihoodComputationMPITP assume that each process will enter the likelihood computation code.
2. A serial likelihood class should be used with serial control only. A serial likelihood computation is only entered by the process which calls it. If MPI is used, serial control will perform the image in parallel.

void BIE::PosteriorProb::Image

(

string

filename

)

Write out an image cut of the posterior.

Parameters:

filename

is image in GNUPLOT format

The user will be prompted for the fixed parameters and range of the two parameters in the cutting plane.

Cuts including mixture values work as follows. If sum of varying weights exceeds 1, these values are normalized and the remaining weights are set to zero. Otherwise, the remaining weight is divided among the fixed weights according to their relative values. For example, a mixture of two components with one varying weight, a, in [0, 1] and one fixed weight, b, of any value, will always be assigned b = 1 - a.

void BIE::PosteriorProb::Image	(	string	filename,
		string	input
	)

Write out an image cut of the posterior.

Parameters:

	filename	is image in GNUPLOT format
	input	is the list of values beginning with weights and followed by parameters for each element in the mixture. The first field is 'f' or 'r' denotied a single fixed value or a range pair to follow.

---

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

• /home/weinberg/src/BIE/include/PosteriorProb.h