Module es.upm.fi.cig.multictbnc

Package es.upm.fi.cig.multictbnc.learning.parameters.ctbn

Class CTBNParameterLearningAlgorithm

java.lang.Object

es.upm.fi.cig.multictbnc.learning.parameters.ctbn.CTBNParameterLearningAlgorithm

All Implemented Interfaces:

ParameterLearningAlgorithm

Direct Known Subclasses:

CTBNBayesianEstimation, CTBNMaximumLikelihoodEstimation

public abstract class CTBNParameterLearningAlgorithm
extends Object
implements ParameterLearningAlgorithm

Define methods for parameter learning algorithms of continuous-time Bayesian networks.

Constructor Summary

Constructors

Constructor	Description
CTBNParameterLearningAlgorithm()

Method Summary

Modifier and Type	Method	Description
protected void	estimateParameters(CIMNode node)	Estimates the parameters for a given node from its computed sufficient statistics.
protected abstract CTBNSufficientStatistics	getSufficientStatisticsNode(DiscreteStateNode node, Dataset dataset)	Returns the sufficient statistics of a DiscreteNode for a given Dataset.
void	learn(Node node, Dataset dataset)	Learns the parameters of a certain node of a PGM.
void	learn(List<? extends Node> nodes, Dataset dataset)	Learns the parameters of a certain PGM.
void	setSufficientStatistics(Node node, Dataset dataset)	Obtains the sufficient statistics of a BN node.
void	setSufficientStatistics(List<? extends Node> nodes, Dataset dataset)	Obtains the sufficient statistics of each node of a CTBN.

Methods inherited from class java.lang.Object

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

Methods inherited from interface es.upm.fi.cig.multictbnc.learning.parameters.ParameterLearningAlgorithm

getIdentifier, getNameMethod, getParametersAlgorithm

Constructor Details

CTBNParameterLearningAlgorithm

public CTBNParameterLearningAlgorithm()

Method Details

setSufficientStatistics

public void setSufficientStatistics(List<? extends Node> nodes,
 Dataset dataset)

Obtains the sufficient statistics of each node of a CTBN.

Parameters:

nodes - list of nodes

dataset - dataset used to compute the sufficient statistics

learn

public void learn(Node node,
 Dataset dataset)

Description copied from interface: ParameterLearningAlgorithm

Learns the parameters of a certain node of a PGM.

Specified by:

learn in interface ParameterLearningAlgorithm

Parameters:

node - node of the PGM

dataset - dataset used to learn the parameters

learn

public void learn(List<? extends Node> nodes,
 Dataset dataset)

Description copied from interface: ParameterLearningAlgorithm

Learns the parameters of a certain PGM.

Specified by:

learn in interface ParameterLearningAlgorithm

Parameters:

nodes - nodes of the PGM

dataset - dataset used to learn the parameters

setSufficientStatistics

public void setSufficientStatistics(Node node,
 Dataset dataset)

Description copied from interface: ParameterLearningAlgorithm

Obtains the sufficient statistics of a BN node.

Specified by:

setSufficientStatistics in interface ParameterLearningAlgorithm

Parameters:

node - node whose sufficient statistics will be computed

dataset - dataset used to compute the sufficient statistics

estimateParameters

protected void estimateParameters(CIMNode node)

Estimates the parameters for a given node from its computed sufficient statistics. Two parameters that summarise the CIMs of each variable are estimated. The first one contains the probabilities of the variables leaving a state and the second one the probabilities of leaving a state for a certain one.

Parameters:

node - node with sufficient statistics where it is stored the computed parameters

getSufficientStatisticsNode

protected abstract CTBNSufficientStatistics getSufficientStatisticsNode(DiscreteStateNode node,
 Dataset dataset)

Returns the sufficient statistics of a DiscreteNode for a given Dataset.

Parameters:

node - a DiscreteNode

dataset - dataset from which the sufficient statistics are extracted

Returns:

sufficient statistics of the provided node