com.mockturtlesolutions.snifflib.invprobs

Class MCMC

java.lang.Object

com.mockturtlesolutions.snifflib.invprobs.MCMC

All Implemented Interfaces:

DefaultReportable, Reportable, ObjectSampler, Optimization, java.lang.Runnable

public class MCMC
extends java.lang.Object
implements DefaultReportable, java.lang.Runnable, ObjectSampler, Optimization

Constructor Summary

Constructors

Constructor and Description
MCMC(MCMCSampler s)

Method Summary

Methods

Modifier and Type	Method and Description
void	addBlock(MCMCBlock newblock) Adds a block of parameters for which cycle interations of MCMC will be performed.
void	addObjectToSample(java.lang.String s)
void	addOptimizationListener(OptimizationListener L)
int	getAcceptCount()
int	getBurnIn(int n)
java.util.LinkedList	getChain()
DefaultReporter	getDefaultReporter(java.lang.String key)
java.util.Set	getDefaultReporters()
DblParamSet	getEstimate() Returns the average of each parameter taken over the chain.
DblParamSet	getEstimate(int N) Returns the average of each parameter taken over the last N elements of the chain.
DblParamSet	getInitialGuess()
int	getMaxIterations(int n)
java.util.LinkedList	getObjectChain() Returns a chain of sampled objects.
java.lang.String[]	getObjectsToSample()
int	getRejectCount()
DblParamSet	getVariance() Returns the variance of each parameter taken over the chain.
void	haltSolution(boolean x)
boolean	isSamplingObject(java.lang.String s)
boolean	isStillRunning()
void	removeAllBlocks()
void	removeBlock(MCMCBlock rmblock)
void	removeObjectToSample(java.lang.String s)
void	removeOptimizationListener(OptimizationListener L)
void	run()
void	setBurnIn(int n)
void	setChainLength(int x)
void	setInitialGuess(DblParamSet guess)
void	setMaxIterations(int n)
void	setObjectsToSample(java.lang.String[] S)
void	solve()

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

MCMC

public MCMC(MCMCSampler s)

Method Detail

addOptimizationListener

public void addOptimizationListener(OptimizationListener L)

Specified by:

addOptimizationListener in interface Optimization

removeOptimizationListener

public void removeOptimizationListener(OptimizationListener L)

Specified by:

removeOptimizationListener in interface Optimization

addObjectToSample

public void addObjectToSample(java.lang.String s)

Specified by:

addObjectToSample in interface ObjectSampler

removeObjectToSample

public void removeObjectToSample(java.lang.String s)

Specified by:

removeObjectToSample in interface ObjectSampler

getObjectsToSample

public java.lang.String[] getObjectsToSample()

Specified by:

getObjectsToSample in interface ObjectSampler

isSamplingObject

public boolean isSamplingObject(java.lang.String s)

Specified by:

isSamplingObject in interface ObjectSampler

setObjectsToSample

public void setObjectsToSample(java.lang.String[] S)

Specified by:

setObjectsToSample in interface ObjectSampler

haltSolution

public void haltSolution(boolean x)

getDefaultReporter

public DefaultReporter getDefaultReporter(java.lang.String key)

Specified by:

getDefaultReporter in interface DefaultReportable

getDefaultReporters

public java.util.Set getDefaultReporters()

Specified by:

getDefaultReporters in interface DefaultReportable

setInitialGuess

public void setInitialGuess(DblParamSet guess)

getInitialGuess

public DblParamSet getInitialGuess()

getChain

public java.util.LinkedList getChain()

getObjectChain

public java.util.LinkedList getObjectChain()

Returns a chain of sampled objects. The content and ordering is coincident with the parameter sample chain obtained by the method getChain(). Typically, although not necessarily, a "link" in the object chain will be a function of the corresponding "link" in the parameter chain.

setChainLength

public void setChainLength(int x)

getEstimate

public DblParamSet getEstimate()

Returns the average of each parameter taken over the chain.

getEstimate

public DblParamSet getEstimate(int N)

Returns the average of each parameter taken over the last N elements of the chain.

getVariance

public DblParamSet getVariance()

Returns the variance of each parameter taken over the chain.

addBlock

public void addBlock(MCMCBlock newblock)

Adds a block of parameters for which cycle interations of MCMC will be performed. When a block cycles are complete the next block will be cycled. An iteration of MCMC is complete when all blocks have been cycled. The process is restarted until convergence is obtained or the maximum number of iterations is reached. MCMC will converge faster when parameters are less correlated. Placing uncorrelated parameters within a block can increase mixing and speed convergence. Typically however there is little general guidance about how to block. Knowledge of the model is usually helpful.

setBurnIn

public void setBurnIn(int n)

setMaxIterations

public void setMaxIterations(int n)

getBurnIn

public int getBurnIn(int n)

getMaxIterations

public int getMaxIterations(int n)

removeBlock

public void removeBlock(MCMCBlock rmblock)

removeAllBlocks

public void removeAllBlocks()

getRejectCount

public int getRejectCount()

getAcceptCount

public int getAcceptCount()

run

public void run()

Specified by:

run in interface java.lang.Runnable

isStillRunning

public boolean isStillRunning()

solve

public void solve()
           throws SnifflibGraphicsException

Throws:

SnifflibGraphicsException