Module es.upm.fi.cig.multictbnc

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

Class BNSufficientStatistics

java.lang.Object

es.upm.fi.cig.multictbnc.learning.parameters.bn.BNSufficientStatistics

All Implemented Interfaces:

SufficientStatistics

public class BNSufficientStatistics
extends Object
implements SufficientStatistics

Compute and store the sufficient statistics of a discrete BN node. The sufficient statistics are:

(1) Nx: number of times a variable takes a certain state.

Constructor Summary

Constructors

Constructor	Description
BNSufficientStatistics(double nHP)	Constructs a BNSufficientStatistics by receiving the hyperparameter of the Dirichlet prior distribution.

Method Summary

Modifier and Type	Method	Description
void	computeSufficientStatistics(DiscreteStateNode node, Dataset dataset)	Computes the sufficient statistics of a node in a BN.
double[][]	getNx()	Returns the sufficient statistics of the node.
double	getNxHyperparameter()	Returns the hyperparameter value of the Dirichlet prior distribution.

Methods inherited from class java.lang.Object

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

Constructor Details

BNSufficientStatistics

public BNSufficientStatistics(double nHP)

Constructs a BNSufficientStatistics by receiving the hyperparameter of the Dirichlet prior distribution.

Parameters:

nHP - number of times the variables are in a certain state while its parents take a certain instantiation (hyperparameter)

Method Details

getNx

public double[][] getNx()

Returns the sufficient statistics of the node. This is the number of times the variable is in a certain state while its parents take a certain instantiation.

Returns:

array with the number of appearances of each state of the node given an instantiation of the parents

getNxHyperparameter

public double getNxHyperparameter()

Returns the hyperparameter value of the Dirichlet prior distribution.

Returns:

hyperparameter value

computeSufficientStatistics

public void computeSufficientStatistics(DiscreteStateNode node,
 Dataset dataset)

Computes the sufficient statistics of a node in a BN. This is the number of times the variable is in a certain state while its parents take a certain instantiation.

Specified by:

computeSufficientStatistics in interface SufficientStatistics

Parameters:

dataset - dataset from which the sufficient statistics are extracted

node - node whose sufficient statistics are computed