stochasticsampling
Class AdaptiveAndSelfIS

java.lang.Object
  |
  +--stochasticsampling.ImportanceSampling
        |
        +--stochasticsampling.AdaptiveAndSelfIS

All Implemented Interfaces:

ImportanceInterface

public class AdaptiveAndSelfIS

extends ImportanceSampling

Title: AdaptiveAndSelfIS.java Description: Core Bayesian Network class Copyright: Copyright (c) 2001 Company: KSU / KDD

Field Summary

boolean	AIS
boolean	changeProbs
java.util.Vector[]	ICPT
boolean	old
boolean	SIS

Constructor Summary

AdaptiveAndSelfIS(BBN nodes, int numsamples, java.lang.String evfile)
AdaptiveAndSelfIS is the constructor and initializes variables

AdaptiveAndSelfIS(BBN nodes, int numsamples, java.lang.String evfile, java.lang.String[][] trainingData)
AdaptiveAndSelfIS is the constructor and initializes variables

Method Summary

void	generateAllSamples() generateAllSamples generates m samples using AIS or SIS
void	generateAllSamples(java.lang.String printingsFile) generateAllSamples generates m samples using AIS or SIS
double	getFitness() getFitness returns the fitness value after sampling (1 - RMSE w/ final ICPTs)
static void	main(java.lang.String[] args)
void	setBBN(BBN newNetwork) setBBN sets the Bayesian network and initializes variables, including the NodeManager

Methods inherited from class stochasticsampling.ImportanceSampling

generateSample, generateSample, reorder

Methods inherited from class java.lang.Object

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

Field Detail

ICPT

public java.util.Vector[] ICPT

AIS

public boolean AIS

SIS

public boolean SIS

old

public boolean old

changeProbs

public boolean changeProbs

Constructor Detail

AdaptiveAndSelfIS

public AdaptiveAndSelfIS(BBN nodes,
                         int numsamples,
                         java.lang.String evfile,
                         java.lang.String[][] trainingData)
                  throws java.io.IOException

AdaptiveAndSelfIS is the constructor and initializes variables

Parameters:

nodes - - the Bayesian network

numsamples - - the number of samples to be taken

evfile - - the evidence file name for the network

trainingData - - a 2D string array of sample instantiations, used to compute exactprobs when using GASLEAK

AdaptiveAndSelfIS

public AdaptiveAndSelfIS(BBN nodes,
                         int numsamples,
                         java.lang.String evfile)
                  throws java.io.IOException

AdaptiveAndSelfIS is the constructor and initializes variables

Parameters:

nodes - - the Bayesian network

totalsamples - - the number of samples to be taken

evfile - - the evidence file name for the network

Method Detail

main

public static void main(java.lang.String[] args)

setBBN

public void setBBN(BBN newNetwork)

setBBN sets the Bayesian network and initializes variables, including the NodeManager

Parameters:

newNetwork - - the Bayesian network to be set

generateAllSamples

public void generateAllSamples(java.lang.String printingsFile)
                        throws java.io.IOException

generateAllSamples generates m samples using AIS or SIS

Parameters:

printingsFile - - the output file for RMSE results

generateAllSamples

public void generateAllSamples()
                        throws java.io.IOException

generateAllSamples generates m samples using AIS or SIS

Overrides:

generateAllSamples in class ImportanceSampling

getFitness

public double getFitness()

getFitness returns the fitness value after sampling (1 - RMSE w/ final ICPTs)

Returns:

double - the fitness value