jbnc.inducers
Class BayesianInducer

java.lang.Object
  jbnc.inducers.BayesianInducer

Direct Known Subclasses:

DiscardingBayesianInducer, FANInducer, NaiveBayesInducer, TANInducer

public abstract class BayesianInducer

extends java.lang.Object

Abstract class for Bayesian network inducers.

Since:

June 1, 1999

Author:

Jarek Sacha

Nested Class Summary

protected class

BayesianInducer.NetLayout Network layout.

Field Summary

protected boolean	debugMode
protected static double	DEFAULT_APLHA_IJK
protected BayesianInducer.NetLayout	layout
protected BayesianNetworks.BayesNet	net Bayesian network representing the classifier.
protected QualityMeasure	qualityMeasure Some Bayesian network inducers may use a quality measure to select the most optimal networ architecture for a given training data set.

Constructor Summary

BayesianInducer()

Method Summary

protected void	clean() Clean member variables.
boolean	getDebugMode() Gets the debug mode flag.
BayesianNetworks.BayesNet	getNetwork() Gets Bayesian network representing the classifier.
QualityMeasure	getQualityMeasure() Gets the current quality measure.
void	reset() Reset the inducer.
void	setDebugMode(boolean mode) Sets the debug mode flag.
void	setQualityMeasure(QualityMeasure qualityMeasure) Sets new quality measure.
void	train(FrequencyCalc fc) Construct bayesian network from the training data
abstract void	train(FrequencyCalc fc, boolean usePriors, double alpha_ijk) Construct bayesian network from the training data
protected void	verifyDataset(DatasetInt dataset) Check validity of the dataset.

Methods inherited from class java.lang.Object

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

Field Detail

DEFAULT_APLHA_IJK

protected static final double DEFAULT_APLHA_IJK

See Also:

Constant Field Values

net

protected BayesianNetworks.BayesNet net

Bayesian network representing the classifier.

layout

protected BayesianInducer.NetLayout layout

debugMode

protected boolean debugMode

qualityMeasure

protected QualityMeasure qualityMeasure

Some Bayesian network inducers may use a quality measure to select the most optimal networ architecture for a given training data set.

Constructor Detail

BayesianInducer

public BayesianInducer()

Method Detail

setDebugMode

public void setDebugMode(boolean mode)

Sets the debug mode flag.

Parameters:

mode - The new DebugMode value

setQualityMeasure

public void setQualityMeasure(QualityMeasure qualityMeasure)

Sets new quality measure.

Parameters:

qualityMeasure - The new QualityMeasure value

getNetwork

public BayesianNetworks.BayesNet getNetwork()
                                     throws java.lang.Exception

Gets Bayesian network representing the classifier.

Returns:

The Network value

Throws:

java.lang.Exception - Description of Exception

getDebugMode

public boolean getDebugMode()

Gets the debug mode flag.

Returns:

The DebugMode value

getQualityMeasure

public QualityMeasure getQualityMeasure()

Gets the current quality measure.

Returns:

The QualityMeasure value

reset

public void reset()

Reset the inducer. Get ready for learning of a new network.

train

public void train(FrequencyCalc fc)
           throws java.lang.Exception

Construct bayesian network from the training data

Parameters:

fc - Description of Parameter

Throws:

java.lang.Exception - Description of Exception

train

public abstract void train(FrequencyCalc fc,
                           boolean usePriors,
                           double alpha_ijk)
                    throws java.lang.Exception

Construct bayesian network from the training data

Parameters:

fc - Description of Parameter

usePriors - Description of Parameter

alpha_ijk - Description of Parameter

Throws:

java.lang.Exception - Description of Exception

clean

protected void clean()

Clean member variables.

verifyDataset

protected void verifyDataset(DatasetInt dataset)
                      throws java.lang.Exception

Check validity of the dataset.

Parameters:

dataset - Description of Parameter

Throws:

java.lang.Exception - Description of Exception