Package com.bayesserver.learning.parameters

Class ParameterLearningProgressInfo

java.lang.Object

com.bayesserver.learning.parameters.ParameterLearningProgressInfo

All Implemented Interfaces:

Cloneable

public final class ParameterLearningProgressInfo
extends Object
implements Cloneable

Provides progress information during ParameterLearning.learn(com.bayesserver.data.EvidenceReaderCommand, com.bayesserver.learning.parameters.ParameterLearningOptions).

Method Summary

Modifier and Type	Method	Description
Double	getDelta()	Gets the relative change in parameters used to determine convergence.
int	getIterationCount()	Gets the current iteration count.
Double	getLogLikelihood()	Gets the current log likelihood value, if calculated
Distribution	getMonitoredDistribution(Node node)	Gets a copy of the current distribution assigned to the [node].
Distribution	getMonitoredDistribution(Node node, NodeDistributionKey key)	Gets a copy of the current distribution assigned to the [node] at the requested order.
Distribution	getMonitoredDistribution(Node node, Integer order)	Gets a copy of the current distribution assigned to the [node] at the requested order.

Methods inherited from class java.lang.Object

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

Method Detail

getMonitoredDistribution

public Distribution getMonitoredDistribution(Node node)

Gets a copy of the current distribution assigned to the [node].

Parameters:

node - The node.

Returns:

A copy of the current node distribution.

getMonitoredDistribution

public Distribution getMonitoredDistribution(Node node,
                                             Integer order)

Gets a copy of the current distribution assigned to the [node] at the requested order.

Parameters:

order - The order of the distribution to retrieve. For temporal nodes that may have more than one distribution.

node - The node.

Returns:

A copy of the current node distribution at a specific order.

getMonitoredDistribution

public Distribution getMonitoredDistribution(Node node,
                                             NodeDistributionKey key)

Gets a copy of the current distribution assigned to the [node] at the requested order.

Parameters:

key - The order/related node of the distribution to retrieve. For temporal nodes / noisy nodes that may have more than one distribution.

node - The node.

Returns:

A copy of the current node distribution at a specific order.

getLogLikelihood

public Double getLogLikelihood()

Gets the current log likelihood value, if calculated

This value is only calculated if ParameterLearningOptions.getMonitorLogLikelihood() is true, as it can be expensive to calculate.

For more information about the log likelihood statistic see ParameterLearningOutput.getLogLikelihood().

Returns:

The log likelihood.

getDelta

public Double getDelta()

Gets the relative change in parameters used to determine convergence. See ParameterLearningOptions.getTolerance() for more information about convergence.

getIterationCount

public int getIterationCount()

Gets the current iteration count.

Returns:

The iteration count.