com.ewasystems.dm.neural.nn
Class NeuralNetworkModel

java.lang.Object
  |
  +--com.ewasystems.dm.AbstractEngineModel
        |
        +--com.ewasystems.dm.neural.nn.NeuralNetworkModel

All Implemented Interfaces:

java.lang.Cloneable, com.ewasystems.dm.util.DirectedAcyclicGraph, java.io.Serializable

public class NeuralNetworkModel

extends com.ewasystems.dm.AbstractEngineModel

implements com.ewasystems.dm.util.DirectedAcyclicGraph

Title: NeuralNetworkModel

Description: A Neural Network Model Containing Neuron DAG

Copyright: (c) 1998 - 2005

Company: EWA Systems

See Also:

Serialized Form

Constructor Summary

NeuralNetworkModel()
Constructs an Empty Neural Network Model

NeuralNetworkModel(int[] layerSizes, int layerType)
Constructs a Neural Network Model

NeuralNetworkModel(int[] layerSizes, int[] layerTypes)
Constructs an Neural Network Model

Method Summary

NeuralLayer	createLayer(int size, int type) Create a Neural Layer
Neuron	createNeuron(int type) Creates a New Neuron
void	destroyLayer(NeuralLayer layer) Disconnects and Destroys a Neural Layer
void	destroyNeuron(Neuron neuron) Disconnects and Destroys a Neuron
java.util.HashSet	getAllDAGNodes() Gets all of the Neurons
com.ewasystems.dm.AbstractEngineConfig	getConfig() Gets the Engine Configuation
NeuralLayer	getCyclicSourceLayer() Returns the First Layer in the Network (Usually the Source Layer)
NeuralLayer	getCyclicTargetLayer() Returns the Last Layer in the Network (Usually the Target Layer)
double	getDataMSE() Gets the Data Mean Squared Error
int	getLayerCount() Gets the Number of Hidden Network Layers
java.util.Iterator	getLayerIterator() Gets the Hidden Layer Iterator
com.ewasystems.util.ExtensionFileFilter	getModelFileFilter() Gets the Model File Filter
int	getNeuronCount() Returns the Total Number of Neurons in the Network
java.util.Iterator	getNeuronIterator() Gets an Neuron Iterator
NeuralLayer	getSourceLayer() Returns the First Layer in the Network (Usually the Source Layer)
NeuralLayer	getTargetLayer() Returns the Last Layer in the Network (Usually the Target Layer)
double	getTestMSE() Gets the Test Mean Squared Error
void	initializeNetwork() Initializes the Network By Setting Random Weights on All Neurons
void	setConfig(com.ewasystems.dm.AbstractEngineConfig config) Sets the Engine Configuation
void	setCyclicSourceLayer(NeuralLayer cyclicSourceLayer) Returns the First Layer in the Network (Usually the Source Layer)
void	setCyclicTargetLayer(NeuralLayer cyclicTargetLayer) Returns the First Layer in the Network (Usually the Source Layer)
void	setSourceLayer(NeuralLayer sourceLayer) Returns the First Layer in the Network (Usually the Source Layer)
void	setTargetLayer(NeuralLayer targetLayer) Returns the First Layer in the Network (Usually the Source Layer)
boolean	validateModel() Validates the DAG

Methods inherited from class com.ewasystems.dm.AbstractEngineModel

loadModel, saveModel

Methods inherited from class java.lang.Object

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

Constructor Detail

NeuralNetworkModel

public NeuralNetworkModel()

Constructs an Empty Neural Network Model

NeuralNetworkModel

public NeuralNetworkModel(int[] layerSizes,
                          int layerType)

Constructs a Neural Network Model

Parameters:

layerSizes - The Layer Sizes

layerType - The Neuron Type

NeuralNetworkModel

public NeuralNetworkModel(int[] layerSizes,
                          int[] layerTypes)

Constructs an Neural Network Model

Parameters:

layerSizes - The Layer Sizes

layerTypes - The Neuron Types

Method Detail

createLayer

public NeuralLayer createLayer(int size,
                               int type)

Create a Neural Layer

Parameters:

size - The Number of Neurons in the Layer

type - The Type of Neurons in the Layer

Returns:

The Neural Layer

createNeuron

public Neuron createNeuron(int type)

Creates a New Neuron

Parameters:

type - The Neuron Type

Returns:

The Neuron

destroyLayer

public void destroyLayer(NeuralLayer layer)

Disconnects and Destroys a Neural Layer

Parameters:

layer - The Neural Layer

destroyNeuron

public void destroyNeuron(Neuron neuron)

Disconnects and Destroys a Neuron

Parameters:

neuron - The Neuron

getAllDAGNodes

public java.util.HashSet getAllDAGNodes()

Gets all of the Neurons

Specified by:

getAllDAGNodes in interface com.ewasystems.dm.util.DirectedAcyclicGraph

Returns:

HashSet: The Set of DAG Nodes

getConfig

public com.ewasystems.dm.AbstractEngineConfig getConfig()

Gets the Engine Configuation

Overrides:

getConfig in class com.ewasystems.dm.AbstractEngineModel

getCyclicSourceLayer

public NeuralLayer getCyclicSourceLayer()

Returns the First Layer in the Network (Usually the Source Layer)

getCyclicTargetLayer

public NeuralLayer getCyclicTargetLayer()

Returns the Last Layer in the Network (Usually the Target Layer)

getDataMSE

public double getDataMSE()

Gets the Data Mean Squared Error

Returns:

The Data MSE

getLayerCount

public int getLayerCount()

Gets the Number of Hidden Network Layers

getLayerIterator

public java.util.Iterator getLayerIterator()

Gets the Hidden Layer Iterator

Returns:

Iterator: The Layer Iterator

getModelFileFilter

public com.ewasystems.util.ExtensionFileFilter getModelFileFilter()

Gets the Model File Filter

Overrides:

getModelFileFilter in class com.ewasystems.dm.AbstractEngineModel

Returns:

ExtensionFileFilter The Model File Filter

getNeuronCount

public int getNeuronCount()

Returns the Total Number of Neurons in the Network

getNeuronIterator

public java.util.Iterator getNeuronIterator()

Gets an Neuron Iterator

Returns:

The Iterator

getSourceLayer

public NeuralLayer getSourceLayer()

Returns the First Layer in the Network (Usually the Source Layer)

getTargetLayer

public NeuralLayer getTargetLayer()

Returns the Last Layer in the Network (Usually the Target Layer)

getTestMSE

public double getTestMSE()

Gets the Test Mean Squared Error

Returns:

The Test MSE

initializeNetwork

public void initializeNetwork()

Initializes the Network By Setting Random Weights on All Neurons

setConfig

public void setConfig(com.ewasystems.dm.AbstractEngineConfig config)

Sets the Engine Configuation

Parameters:

config - The Engine Configuration

setCyclicSourceLayer

public void setCyclicSourceLayer(NeuralLayer cyclicSourceLayer)

Returns the First Layer in the Network (Usually the Source Layer)

Parameters:

cyclicSourceLayer - The First Neural Layer

setCyclicTargetLayer

public void setCyclicTargetLayer(NeuralLayer cyclicTargetLayer)

Returns the First Layer in the Network (Usually the Source Layer)

Parameters:

cyclicTargetLayer - The First Neural Layer

setSourceLayer

public void setSourceLayer(NeuralLayer sourceLayer)

Returns the First Layer in the Network (Usually the Source Layer)

Parameters:

sourceLayer - The First Neural Layer

setTargetLayer

public void setTargetLayer(NeuralLayer targetLayer)

Returns the First Layer in the Network (Usually the Source Layer)

Parameters:

targetLayer - The First Neural Layer

validateModel

public boolean validateModel()

Validates the DAG

Returns:

True if Valid