yawn.nn.appart
Class AppArt

java.lang.Object
  yawn.nn.NeuralNetwork
      yawn.nn.appart.AppArt

All Implemented Interfaces:

java.io.Serializable

Direct Known Subclasses:

GasArt

public class AppArt
extends NeuralNetwork

This class implements the AppART neural network. See:

Martí, L., Policriti, A. & García, L. (2003). Hybrid Adaptive Resonance Theory Neural Networks for Universal Function Approximation. In Abraham, A. and Jain, L. (eds.), Innovations in Intelligent Systems and Applications, Studies in Fuzziness and Soft Computing Series. Heidelberg: Physica (Springer) Verlag.

Version:

$Revision: 1.13 $

Author:

Luis Martí (luis dot marti at uc3m dot es)

See Also:

Serialized Form

Field Summary

protected double	desiredMeanSquaredError
protected GainControlUnitOnMatching	f2GainControl
protected RecognitionLayer	f2Layer
protected Pattern	initialDeviations
protected InputLayer	inputLayer
protected double	learningRate
protected boolean	matchTrackingOneShot
protected long	maxEpochs
protected GainControlUnitOnOutput	outputGainControl
protected OutputLayer	outputLayer
protected double	predictionError
protected PredictionLayer	predictionLayer
protected double	predictionLayerLearningRate
protected double	testMatchVigilance
protected double	trainMatchVigilance
protected boolean	useAbsoluteError

Constructor Summary

AppArt()

Method Summary

protected void	computeError(Pattern prediction, Pattern expected)
protected void	doBackTrack()
protected void	doMatchTracking()
double	getDesiredMeanSquaredError()
RecognitionLayer	getF2Layer()
Pattern	getInitialDeviations()
int	getInputSize()
double	getLearningRate()
long	getMaxEpochs()
java.lang.String	getNeuralNetworkName() Returns a human readable
int	getOutputSize()
double	getPredictionError()
PredictionLayer	getPredictionLayer()
double	getPredictionLayerLearningRate()
double	getTestMatchVigilance()
double	getTrainMatchVigilance()
protected void	init()
boolean	isMatchTrackingOneShot()
boolean	isUseAbsoluteError()
protected void	learn(Pattern pat)
protected Pattern	learnNewCategory(Pattern output)
protected int	numberOfF2Nodes()
void	oneLearningStep(Pattern input, Pattern output) Performs one learning iteration.
Pattern	predict(Pattern input) Computes a network prediction
protected Pattern	propagate(Pattern output) Propagates the input already already presented to the network (by calling setInput())
protected void	reset()
void	setAdapting(boolean adapt) If set to true puts the network in adapting (training) mode; false puts the network in static or frozen (testing) mode.
void	setDesiredMeanSquaredError(double desiredMeanSquaredError)
void	setF2GainControl(GainControlUnitOnMatching gf2)
void	setF2Layer(RecognitionLayer f2)
void	setInitialDeviations(Pattern initialDeviations)
void	setInput(Pattern input)
void	setLearningRate(double learningRate)
void	setMatchTrackingOneShot(boolean matchTrackingOneShot)
void	setMaxEpochs(long maxEpochs)
void	setPredictionError(double predictionError)
void	setPredictionLayer(PredictionLayer predictionLayer)
void	setPredictionLayerLearningRate(double predictionLayerLearningRate)
void	setTestMatchVigilance(double testMatchVigilance)
void	setTrainMatchVigilance(double trainMatchVigilance)
void	setup(NeuralNetworkConfig config)
void	setUseAbsoluteError(boolean error)
void	setUseRelativeError(boolean error)
void	train(InputOutputPattern[] iop) Trains the network until the stop criteria is met.
void	useMinimumActivationMatchTracking(boolean use)
void	useOneShotMatchTracking(boolean use)
NeuralNetworkConfig	yieldConfiguration() (non-Javadoc)

Methods inherited from class yawn.nn.NeuralNetwork

getAdapting, getStatisticsFacility, isAdapting, setStatisticsFacility

Methods inherited from class java.lang.Object

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

Field Detail

f2GainControl

protected GainControlUnitOnMatching f2GainControl

f2Layer

protected RecognitionLayer f2Layer

initialDeviations

protected Pattern initialDeviations

inputLayer

protected InputLayer inputLayer

learningRate

protected double learningRate

matchTrackingOneShot

protected boolean matchTrackingOneShot

maxEpochs

protected long maxEpochs

outputGainControl

protected GainControlUnitOnOutput outputGainControl

outputLayer

protected OutputLayer outputLayer

predictionError

protected double predictionError

predictionLayer

protected PredictionLayer predictionLayer

predictionLayerLearningRate

protected double predictionLayerLearningRate

testMatchVigilance

protected double testMatchVigilance

trainMatchVigilance

protected double trainMatchVigilance

useAbsoluteError

protected boolean useAbsoluteError

desiredMeanSquaredError

protected double desiredMeanSquaredError

Constructor Detail

AppArt

public AppArt()

Method Detail

computeError

protected void computeError(Pattern prediction,
                            Pattern expected)

doBackTrack

protected void doBackTrack()

doMatchTracking

protected void doMatchTracking()

getF2Layer

public RecognitionLayer getF2Layer()

Returns:

Returns the f2Layer.

getInitialDeviations

public Pattern getInitialDeviations()

Returns:

Returns the initialDeviations.

getInputSize

public int getInputSize()

Specified by:

getInputSize in class NeuralNetwork

getLearningRate

public double getLearningRate()

Returns:

Returns the learningRate.

getMaxEpochs

public long getMaxEpochs()

Returns:

Returns the maxEpochs.

getNeuralNetworkName

public java.lang.String getNeuralNetworkName()

Description copied from class: NeuralNetwork

Returns a human readable

Specified by:

getNeuralNetworkName in class NeuralNetwork

Returns:

The name

See Also:

NeuralNetwork.getNeuralNetworkName()

getOutputSize

public int getOutputSize()

Specified by:

getOutputSize in class NeuralNetwork

getPredictionError

public double getPredictionError()

Returns:

Returns the predictionError.

getPredictionLayer

public PredictionLayer getPredictionLayer()

Returns:

Returns the predictionLayer.

getPredictionLayerLearningRate

public double getPredictionLayerLearningRate()

Returns:

Returns the predictionLayerLearningRate.

getTestMatchVigilance

public double getTestMatchVigilance()

Returns:

Returns the testMatchVigilance.

getTrainMatchVigilance

public double getTrainMatchVigilance()

Returns:

Returns the trainMatchVigilance.

init

protected void init()

isMatchTrackingOneShot

public boolean isMatchTrackingOneShot()

Returns:

Returns the matchTrackingOneShot.

isUseAbsoluteError

public boolean isUseAbsoluteError()

Returns:

Returns the useAbsoluteError.

learn

protected void learn(Pattern pat)

learnNewCategory

protected Pattern learnNewCategory(Pattern output)

numberOfF2Nodes

protected int numberOfF2Nodes()

oneLearningStep

public void oneLearningStep(Pattern input,
                            Pattern output)

Description copied from class: NeuralNetwork

Performs one learning iteration.

Specified by:

oneLearningStep in class NeuralNetwork

Parameters:

input - The network input.

output - The desired output.

predict

public Pattern predict(Pattern input)

Description copied from class: NeuralNetwork

Computes a network prediction

Specified by:

predict in class NeuralNetwork

Parameters:

input - The input to propagate.

Returns:

the network output.

propagate

protected Pattern propagate(Pattern output)

Propagates the input already already presented to the network (by calling setInput())

reset

protected void reset()

setAdapting

public void setAdapting(boolean adapt)

Description copied from class: NeuralNetwork

If set to true puts the network in adapting (training) mode; false puts the network in static or frozen (testing) mode.

Overrides:

setAdapting in class NeuralNetwork

setF2GainControl

public void setF2GainControl(GainControlUnitOnMatching gf2)

Parameters:

gf2 - the f2GainControl to set.

setF2Layer

public void setF2Layer(RecognitionLayer f2)

Parameters:

f2 - The f2Layer to set.

setInitialDeviations

public void setInitialDeviations(Pattern initialDeviations)

Parameters:

initialDeviations - The initialDeviations to set.

setInput

public void setInput(Pattern input)

setLearningRate

public void setLearningRate(double learningRate)

setMatchTrackingOneShot

public void setMatchTrackingOneShot(boolean matchTrackingOneShot)

Parameters:

matchTrackingOneShot - The matchTrackingOneShot to set.

setMaxEpochs

public void setMaxEpochs(long maxEpochs)

Parameters:

maxEpochs - The maxEpochs to set.

setPredictionError

public void setPredictionError(double predictionError)

Parameters:

predictionError - The predictionError to set.

setPredictionLayer

public void setPredictionLayer(PredictionLayer predictionLayer)

Parameters:

predictionLayer - the predictionLayer to set

setPredictionLayerLearningRate

public void setPredictionLayerLearningRate(double predictionLayerLearningRate)

Parameters:

predictionLayerLearningRate - The predictionLayerLearningRate to set.

setTestMatchVigilance

public void setTestMatchVigilance(double testMatchVigilance)

Parameters:

testMatchVigilance - The testMatchVigilance to set.

setTrainMatchVigilance

public void setTrainMatchVigilance(double trainMatchVigilance)

Parameters:

trainMatchVigilance - The trainMatchVigilance to set.

setup

public void setup(NeuralNetworkConfig config)
           throws ConfigurationException

Specified by:

setup in class NeuralNetwork

Throws:

ConfigurationException

setUseAbsoluteError

public void setUseAbsoluteError(boolean error)

setUseRelativeError

public void setUseRelativeError(boolean error)

train

public void train(InputOutputPattern[] iop)

Description copied from class: NeuralNetwork

Trains the network until the stop criteria is met.

Specified by:

train in class NeuralNetwork

Parameters:

iop - The training set to be learned.

useMinimumActivationMatchTracking

public void useMinimumActivationMatchTracking(boolean use)

useOneShotMatchTracking

public void useOneShotMatchTracking(boolean use)

yieldConfiguration

public NeuralNetworkConfig yieldConfiguration()

(non-Javadoc)

Specified by:

yieldConfiguration in class NeuralNetwork

See Also:

NeuralNetwork.yieldConfiguration()

getDesiredMeanSquaredError

public double getDesiredMeanSquaredError()

Returns:

Returns the desiredMeanSquaredError.

setDesiredMeanSquaredError

public void setDesiredMeanSquaredError(double desiredMeanSquaredError)

Parameters:

desiredMeanSquaredError - The desiredMeanSquaredError to set.