BIE::PostMixturePrior Class Reference

Prior for mixture model. More...

#include <PostMixturePrior.h>

Inheritance diagram for BIE::PostMixturePrior:

[legend]

Collaboration diagram for BIE::PostMixturePrior:

[legend]

List of all members.

Public Member Functions
virtual void	SamplePrior (int &m, vector< double > &w, vector< vector< double > > &p)
	Sample: return weights and components separately.
virtual void	SamplePrior (int M, int n, vector< double > &V)
	Sample: return a single component vector.
virtual void	SampleProposal (Chain &ch, MHWidth *width)
	Based on prior type, return random variates for mixture model.
virtual int	MinC ()
	State with minimum number of components.
virtual int	MaxC ()
	State with maximum number of components.
Constructors
	PostMixturePrior ()
	For cloning only: makes an uninitialized prior.
	PostMixturePrior (MixturePrior *old, Ensemble *dist)
	The useful constructor (a poor person's multiple inheritance).
	PostMixturePrior (MixturePrior *old, Ensemble *dist, int nmix, int ndim, int level, int nburn, string filename, int keypos)
	Construct an instance using MCMC log file.
	~PostMixturePrior ()
	Destructor.
Distribution members
The packing order in vectors is: Nmix weights, followed by Nmix vectors of Ndim elements
PostMixturePrior *	New ()
	Object factor (clone).
void	FreeRange ()
	Turn off the variable range limit enforcement (on by default).
void	useSamples ()
	Use the difference between a prior sample and the mean for setting the proposal (the default).
void	useWidths ()
	Use the difference standard deviation widths for setting the proposal.
double	PDF (State &x)
	Differential distribution function P(x).
double	logPDF (State &x)
	Log of differential distribution function P(x).
double	logPDFMarginal (int M, int n, vector< double > &x)
	Log of differential distribution function P(x) for component n in subspace M.
vector< double >	lower (void)
	Lower bound on distribution.
vector< double >	upper (void)
	Upper bound on distribution.
vector< double >	Mean ()
	Return mean of distribution.
vector< double >	StdDev ()
	Return standard deviation of distribution.
vector< double >	Moments (int m)
	Return specifided moment of distribution.
vector< double >	Sample ()
	Return random variate from distribution.
vector< double >	Sample (int m)
	Return random variate from distribution.
Static Public Attributes
Global variables
static double	width_factor
	Width factor for Metropolis-Hastings step using PostMixturePrior.
static int	max_iter
	Maximum number of iterations for sampling prior.
Friends
class	boost::serialization::access

---

Detailed Description

Prior for mixture model.

---

Constructor & Destructor Documentation

BIE::PostMixturePrior::PostMixturePrior	(	MixturePrior *	old,
		Ensemble *	dist,
		int	nmix,
		int	ndim,
		int	level,
		int	nburn,
		string	filename,
		int	keypos
	)

Construct an instance using MCMC log file.

This constructor may be used to sample from a particular MCMC run.

---

Member Function Documentation

virtual void BIE::PostMixturePrior::SamplePrior	(	int	M,
		int	n,
		vector< double > &	V
	)			[virtual]

Sample: return a single component vector.

Parameters:

	V	from subspace
	M	and component index
	n

Implements BIE::MixturePrior.

---

Member Data Documentation

double BIE::PostMixturePrior::width_factor [static]

Width factor for Metropolis-Hastings step using PostMixturePrior.

The input width is assumed to be a variance and the sample width is chosen following Gelman et al. 1996, ideal if the underlying posterior were Gaussian (see Efficient Metropolis jumping rules, Bayesian Statistics V, eds. Bernado, Berger, David & Smith, Oxford, pp. 599-608)

---

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

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