jas.ai.ann
Class NeuralDataSet

java.lang.Object
  jas.ai.ann.NeuralDataSet

public class NeuralDataSet
extends java.lang.Object

A datawarehouse for neural networks. It contains data to make ANN learning and testing.

Title: JAS

Description: Java Agent-based Simulation library

Copyright: 2003 Michele Sonnessa

This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307, USA.

Author:

Michele Sonnessa

Field Summary

static java.lang.String

TAG_DS

Constructor Summary

NeuralDataSet(int inputNodes, int outputNodes)
Create a new neural data set with specified parameters.

NeuralDataSet(int inputNodes, int outputNodes, java.util.Random randomizer)
Create a new neural data set with specified parameters.

Method Summary

void	appendCurrentNNSample(NeuralNetwork nn) Append current input/target values of the given ANN to the sample list.
void	appendCurrentNNSample(NeuralNetwork nn, boolean useInTraining) Append current input/target values of the given ANN to the sample list, with the given training attribute.
void	appendSample(double[] inputSet, double[] targetSet)
void	appendSample(double[] inputSet, double[] targetSet, boolean useInTraining) Append data to the sample list.
void	assignCurrentNNSample(NeuralNetwork nn) Assign to given ANN the current sample, according to the state of the iterator.
void	clearDataSet(int inputNodes, int outputNodes) Empty the data set.
double[]	getCurrentInputArray() Return a copy of the current input layer.
int	getCurrentIteratorIndex() Return the current index of the iterator.
double[]	getCurrentTargetArray() Return a copy of the current target layer.
Jama.Matrix	getInputMinMaxMatrix() Return a Jama.Matrix of minimum and maximum values for the input nodes.
Jama.Matrix	getTargetMinMaxMatrix() Return a Jama.Matrix of minimum and maximum values for the target nodes.
boolean	isCurrentSampleInTraining() Tell if the current sample is to be used in learning.
boolean	nextItem() Move the iterator to the next element.
boolean	nextTrainingItem() Move the iterator to the next element, which has the useInTraining attribute set to true.
void	randomizeDataSet() Shuffle the sample list.
void	removeFirst() Remove the first sample from the list.
void	removeFirst(int numberOfSamples) Remove the first given number of samples from the list.
void	resetIterator() Reset the iterator to the first element of the list.
int	size() Return the current size of the sample list.
void	sortDataSetInOriginalSequence() Reset the original sequence of the sample list.
void	XML_appendDataSet(org.w3c.dom.Document document, org.w3c.dom.Node rootNode) Write the whole data set into an XML document structure, below the given root node.
void	XML_extractDataSet(org.w3c.dom.Node rootNode) Read the data set from an XML document structure, below the given root node.

Methods inherited from class java.lang.Object

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

Field Detail

TAG_DS

public static final java.lang.String TAG_DS

See Also:

Constant Field Values

Constructor Detail

NeuralDataSet

public NeuralDataSet(int inputNodes,
                     int outputNodes)

Create a new neural data set with specified parameters.

Parameters:

inputNodes - The number of input nodes in the ANN input layer.

outputNodes - The number of output nodes in the ANN output layer.

NeuralDataSet

public NeuralDataSet(int inputNodes,
                     int outputNodes,
                     java.util.Random randomizer)

Create a new neural data set with specified parameters.

Parameters:

inputNodes - The number of input nodes in the ANN input layer.

outputNodes - The number of output nodes in the ANN output layer.

randomizer - A random generator used to shuffle the sample list.

Method Detail

appendCurrentNNSample

public void appendCurrentNNSample(NeuralNetwork nn)

Append current input/target values of the given ANN to the sample list. The appended samples is given the true useInTraining attribute.

Parameters:

nn - The ANN from which data are taken.

appendCurrentNNSample

public void appendCurrentNNSample(NeuralNetwork nn,
                                  boolean useInTraining)

Append current input/target values of the given ANN to the sample list, with the given training attribute.

Parameters:

nn - The ANN from which data are taken.

useInTraining - If true the sample will be used by the learning algorithm. If false it will be used only during the testing.

appendSample

public void appendSample(double[] inputSet,
                         double[] targetSet)

appendSample

public void appendSample(double[] inputSet,
                         double[] targetSet,
                         boolean useInTraining)

Append data to the sample list.

Parameters:

inputSet - An array representing the input layer.

targetSet - An array representing the output expected layer.

useInTraining - Use in training or not.

Throws:

java.lang.ArrayIndexOutOfBoundsException - If inputSet or targetSet length does not match with input/output length of the ANN.

assignCurrentNNSample

public void assignCurrentNNSample(NeuralNetwork nn)

Assign to given ANN the current sample, according to the state of the iterator.

Parameters:

nn - The target ANN.

clearDataSet

public void clearDataSet(int inputNodes,
                         int outputNodes)

Empty the data set.

Parameters:

inputNodes - The new input layer length.

outputNodes - The new output layer length.

getCurrentInputArray

public double[] getCurrentInputArray()

Return a copy of the current input layer.

Returns:

An array containing a copy the current input layer.

getCurrentIteratorIndex

public int getCurrentIteratorIndex()

Return the current index of the iterator.

Returns:

The current index of the iterator.

getCurrentTargetArray

public double[] getCurrentTargetArray()

Return a copy of the current target layer.

Returns:

An array containing a copy the current target layer.

getInputMinMaxMatrix

public Jama.Matrix getInputMinMaxMatrix()

Return a Jama.Matrix of minimum and maximum values for the input nodes. The row 0 contains min values and the row 1 the max ones.

Returns:

The min-max matrix.

getTargetMinMaxMatrix

public Jama.Matrix getTargetMinMaxMatrix()

Return a Jama.Matrix of minimum and maximum values for the target nodes. The row 0 contains min values and the row 1 the max ones.

Returns:

The min-max matrix.

isCurrentSampleInTraining

public boolean isCurrentSampleInTraining()

Tell if the current sample is to be used in learning.

Returns:

True if current sample is to be used in learning, false otherwise.

nextItem

public boolean nextItem()

Move the iterator to the next element.

Returns:

True if the last element in the list has been not reached yet.

nextTrainingItem

public boolean nextTrainingItem()

Move the iterator to the next element, which has the useInTraining attribute set to true.

Returns:

True if the last element in the list has been not reached yet.

randomizeDataSet

public void randomizeDataSet()

Shuffle the sample list.

removeFirst

public void removeFirst()

Remove the first sample from the list. WARNING: The first value is the current first value. If you call a randomizeDataSet the first element will be no longer the first appended one. To remove it you have to call sortDataSetInOriginalSequence() first.

removeFirst

public void removeFirst(int numberOfSamples)

Remove the first given number of samples from the list. WARNING: The first value is the current first value. If you call a randomizeDataSet the first element will be no longer the first appended one. To remove it you have to call sortDataSetInOriginalSequence() first.

Parameters:

numberOfSamples - The first number of samples to be removed.

resetIterator

public void resetIterator()

Reset the iterator to the first element of the list.

size

public int size()

Return the current size of the sample list.

Returns:

The size of the sample list.

sortDataSetInOriginalSequence

public void sortDataSetInOriginalSequence()

Reset the original sequence of the sample list.

XML_appendDataSet

public void XML_appendDataSet(org.w3c.dom.Document document,
                              org.w3c.dom.Node rootNode)

Write the whole data set into an XML document structure, below the given root node.

Parameters:

document - A well-formed XML document.

rootNode - The root node for the data set structure.

XML_extractDataSet

public void XML_extractDataSet(org.w3c.dom.Node rootNode)

Read the data set from an XML document structure, below the given root node.

Parameters:

rootNode - The root node for the data set structure.