jhpro.nnet

Class BackpropagationNet

java.lang.Object

jhpro.nnet.BackpropagationNet

All Implemented Interfaces:

java.io.Serializable

public class BackpropagationNet
extends java.lang.Object

Backpropagation is a supervised learning algorithm and is mainly used by Multi-Layer-Perceptrons to change the weights connected to the net's hidden neuron layer(s). The backpropagation algorithm uses a computed output error to change the weight values in backward direction. To get this net error, a forwardpropagation phase must have been done before. While propagating in forward direction, the neurons are being activated using the sigmoid activation function.

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor and Description
BackpropagationNet() Construct Backpropagation.

Method Summary

Modifier and Type	Method and Description
void	addNeuronLayer(int i) Adds a neuron layer with size neurons.
void	connectLayers() Connects all neuron layers with weight matrices.
boolean	displayNow()
void	error(int i)
boolean	finishedLearning()
double	getAccuracy() Returns the accuracy value.
java.lang.String	getAsciiValue(java.lang.String s)
java.lang.String	getElapsedTime() Returns the time that elapsed since the learning process started.
double	getError() Returns the current error of the net.
java.lang.String	getInputPattern(int i)
int	getLearningCycle() Returns the current learning cycle of the net.
double	getLearningRate() Returns the current learning rate of the net.
int	getMaxLearningCycles()
double	getMinimumError() Returns the minimum error of a neural net.
float[]	getNeuronOutputs(int i) Returns the output values of all neurons in layer i
int	getNumberOfLayers()
int	getNumberOfNeurons(int i)
int	getNumberOfPatterns()
int	getNumberOfWeights()
int	getNumberOfWeights(int i)
java.lang.String	getOutputPattern(int i) Returns the output pattern with number patternNr.
float	getPatternError(int i) Returns the error of output pattern patternNr.
java.lang.String	getTargetPattern(int i) Returns the target pattern with number patternNr.
float[][]	getWeightValues(int i) Returns the weight values of weight matrix matrixNr.
void	incLearningCycle()
void	learn() Performs one learning step.
void	readConversionFile(java.lang.String s) Reads a conversion table for ASCII-binary values from file conversionFileName.
void	readPatternFile(java.lang.String s)
java.lang.String	recall(java.lang.String s) Tries to recall the correct output for a learned input pattern recallInput.
void	resetTime()
void	setAccuracy(double d) Sets the accuracy value.
void	setDisplayStep(int i)
void	setLearningRate(double d)
void	setMaxLearningCycles(int i) Sets the maximum number of learning cycles to x.
void	setMinimumError(double d) Sets the minimum error of a neural net.
double	square(double d)

Methods inherited from class java.lang.Object

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

Constructor Detail

BackpropagationNet

public BackpropagationNet()

Construct Backpropagation.

Method Detail

addNeuronLayer

public void addNeuronLayer(int i)

Adds a neuron layer with size neurons. Note that neuron layers are sequentially added to the net.

Parameters:

i -

connectLayers

public void connectLayers()

Connects all neuron layers with weight matrices. Must be called after all neuron layers have been added.

setMinimumError

public void setMinimumError(double d)

Sets the minimum error of a neural net.

Parameters:

d -

getMinimumError

public double getMinimumError()

Returns the minimum error of a neural net.

setAccuracy

public void setAccuracy(double d)

Sets the accuracy value.

Parameters:

d -

getAccuracy

public double getAccuracy()

Returns the accuracy value.

Returns:

getWeightValues

public float[][] getWeightValues(int i)

Returns the weight values of weight matrix matrixNr. The values for matrixNr start with zero!

Parameters:

i -

Returns:

getNeuronOutputs

public float[] getNeuronOutputs(int i)

Returns the output values of all neurons in layer i

Parameters:

i - layer

Returns:

getNumberOfLayers

public int getNumberOfLayers()

getNumberOfNeurons

public int getNumberOfNeurons(int i)

getNumberOfWeights

public int getNumberOfWeights()

getNumberOfWeights

public int getNumberOfWeights(int i)

getNumberOfPatterns

public int getNumberOfPatterns()

getInputPattern

public java.lang.String getInputPattern(int i)

getTargetPattern

public java.lang.String getTargetPattern(int i)

Returns the target pattern with number patternNr.

Parameters:

i -

Returns:

getOutputPattern

public java.lang.String getOutputPattern(int i)

Returns the output pattern with number patternNr.

Parameters:

i -

Returns:

getPatternError

public float getPatternError(int i)

Returns the error of output pattern patternNr.

Parameters:

i -

Returns:

getError

public double getError()

Returns the current error of the net.

Returns:

learn

public void learn()

Performs one learning step.

recall

public java.lang.String recall(java.lang.String s)

Tries to recall the correct output for a learned input pattern recallInput.

Parameters:

s -

Returns:

readConversionFile

public void readConversionFile(java.lang.String s)

Reads a conversion table for ASCII-binary values from file conversionFileName.

Parameters:

s -

getAsciiValue

public java.lang.String getAsciiValue(java.lang.String s)

readPatternFile

public void readPatternFile(java.lang.String s)

finishedLearning

public boolean finishedLearning()

error

public void error(int i)

square

public double square(double d)

setLearningRate

public void setLearningRate(double d)

getLearningRate

public double getLearningRate()

Returns the current learning rate of the net.

Returns:

setDisplayStep

public void setDisplayStep(int i)

displayNow

public boolean displayNow()

resetTime

public void resetTime()

getElapsedTime

public java.lang.String getElapsedTime()

Returns the time that elapsed since the learning process started.

Returns:

setMaxLearningCycles

public void setMaxLearningCycles(int i)

Sets the maximum number of learning cycles to x. The default value is -1 (no maximum).

Parameters:

i -

getMaxLearningCycles

public int getMaxLearningCycles()

incLearningCycle

public void incLearningCycle()

getLearningCycle

public int getLearningCycle()

Returns the current learning cycle of the net.

Returns: