lio.probdistributions.bitchain
Class BivariateProbDistribution

java.lang.Object
  lio.probdistributions.bitchain.BinaryProbDistribution
      lio.probdistributions.bitchain.BivariateProbDistribution

All Implemented Interfaces:

LiOResource, ProbDistribution

Direct Known Subclasses:

Chain, Tree

public abstract class BivariateProbDistribution
extends BinaryProbDistribution

This class implements the common features and data structures of bivariate probability distributions. They are separated from bayesian and dependency networks for efficiency. Position of the arrays follows a topological order. The learning method will determine the instantiation of this class.

Field Summary

protected double[][]	conditionals Conditional probabilities of each variable.
protected double[]	marginals Vector that contains the marginal probabilities
protected int[]	parents Variable parent of the one represented by each position.
protected int[]	variables Contains the variable represented by each position.

Fields inherited from class lio.probdistributions.bitchain.BinaryProbDistribution

data, laplace, size

Constructor Summary

BivariateProbDistribution()

Method Summary

protected void	createStructures(int pSize) Creates the necessary structures for the search.
Individual[]	sample(int nIndividuals) Samples some individuals from a probability distribution

Methods inherited from class lio.probdistributions.bitchain.BinaryProbDistribution

isLaplace, setLaplace

Methods inherited from class java.lang.Object

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

Methods inherited from interface lio.probdistributions.ProbDistribution

learn

Methods inherited from interface lio.core.LiOResource

getDefinition

Field Detail

variables

protected int[] variables

Contains the variable represented by each position.

parents

protected int[] parents

Variable parent of the one represented by each position.

conditionals

protected double[][] conditionals

Conditional probabilities of each variable.

marginals

protected double[] marginals

Vector that contains the marginal probabilities

Constructor Detail

BivariateProbDistribution

public BivariateProbDistribution()

Method Detail

createStructures

protected void createStructures(int pSize)

Creates the necessary structures for the search.

Parameters:

pSize - Size of the individuals from which the model will be learnt.

sample

public Individual[] sample(int nIndividuals)

Samples some individuals from a probability distribution

Parameters:

nIndividuals - Number of individuals that must be sampled.

Returns:

The selected individuals.