Overview  Package   Class  Use  Deprecated  Index
Home >> All >> AI >> [ NeuralNetworks overview ]	PREV CLASS  NEXT CLASS
SUMMARY:  JAVADOC |  SOURCE |  DOWNLOAD | NESTED | FIELD | CONSTR | METHOD	DETAIL: FIELD | CONSTR | METHOD

---

AI.NeuralNetworks
Class FeedForwardNetwork  

java.lang.Object
  AI.NeuralNetworks.FeedForwardNetwork

Direct Known Subclasses:

BackErrorPropagation

---

public class FeedForwardNetwork

extends java.lang.Object

This Class is a nerual network with a multiple layer feed forward architecture

---

Field Summary
protected static java.lang.String	COMMENT_TOKEN
protected static java.lang.String	HEXADECIMAL_VERSION
protected static java.lang.String	INPUT_LAYER_TOKEN
protected  java.util.LinkedList	layers
protected static java.lang.String	NUMBER_OF_LAYERS_TOKEN
protected static java.lang.String	VERSION
protected static java.lang.String	VERSION_TOKEN
protected static java.lang.String	WEIGHTED_LAYER_TOKEN
protected static java.lang.String	WEIGHTS_TOKEN

 

Constructor Summary
protected	FeedForwardNetwork()           Creates a new FeedForwardNetwork This constructor should be only used by sublcases
	FeedForwardNetwork(java.io.BufferedReader input)           Creates a new FeedForwardNetwork form the specified BufferedReader
	FeedForwardNetwork(int[] structure)           Creates a new FeedForwardNetwork
	FeedForwardNetwork(int[] netIntArray, int chanchada)           Creates a new FeedForwardNetwork form the specified BufferedReader
	FeedForwardNetwork(java.lang.String fileName)           Creates a new FeedForwardNetwork form the specified file

 

Method Summary
private static int[]	addIntToArray(int[] array, int element, int pos)
private static float[][]	createWeights(java.io.BufferedReader input, int layerSize, int previousLayerSize)
private static java.util.StringTokenizer	getNextStringTokenizer(java.io.BufferedReader input, java.lang.String token)
int	inputSize()           returns the number of inputs of the net without counting the bias
int[]	netToIntArrey()           returns a int[] which represents a netework the weightValues are converted to int with the Float.floatToIntBits() method.
int	outputSize()           returns the number of outputs of the net without counting the bias
float[]	runVector(float[] inputVector)           Runs the network with the inputVector and returns the outputVector
void	save(java.io.BufferedWriter output)           Saves the neural network
void	saveToFile(java.lang.String fileName)           Saves the neural network
java.lang.String	toString()           retrurns a String representation of the net

 

Methods inherited from class java.lang.Object

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

 

Field Detail

layers

protected java.util.LinkedList layers

---

COMMENT_TOKEN

protected static final java.lang.String COMMENT_TOKEN

See Also:

Constant Field Values

---

NUMBER_OF_LAYERS_TOKEN

protected static final java.lang.String NUMBER_OF_LAYERS_TOKEN

See Also:

Constant Field Values

---

INPUT_LAYER_TOKEN

protected static final java.lang.String INPUT_LAYER_TOKEN

See Also:

Constant Field Values

---

WEIGHTED_LAYER_TOKEN

protected static final java.lang.String WEIGHTED_LAYER_TOKEN

See Also:

Constant Field Values

---

WEIGHTS_TOKEN

protected static final java.lang.String WEIGHTS_TOKEN

See Also:

Constant Field Values

---

VERSION_TOKEN

protected static final java.lang.String VERSION_TOKEN

See Also:

Constant Field Values

---

VERSION

protected static final java.lang.String VERSION

See Also:

Constant Field Values

---

HEXADECIMAL_VERSION

protected static final java.lang.String HEXADECIMAL_VERSION

See Also:

Constant Field Values

Constructor Detail

FeedForwardNetwork

protected FeedForwardNetwork()

Creates a new FeedForwardNetwork This constructor should be only used by sublcases

---

FeedForwardNetwork

public FeedForwardNetwork(int[] structure)

Creates a new FeedForwardNetwork

---

FeedForwardNetwork

public FeedForwardNetwork(java.lang.String fileName)
                   throws java.io.IOException

Creates a new FeedForwardNetwork form the specified file

---

FeedForwardNetwork

public FeedForwardNetwork(java.io.BufferedReader input)
                   throws java.io.IOException,
                          java.lang.NumberFormatException

Creates a new FeedForwardNetwork form the specified BufferedReader

---

FeedForwardNetwork

public FeedForwardNetwork(int[] netIntArray,
                          int chanchada)
                   throws java.io.IOException,
                          java.lang.NumberFormatException

Creates a new FeedForwardNetwork form the specified BufferedReader

Method Detail

runVector

public float[] runVector(float[] inputVector)
                  throws UnexpectedInputArraySizeException

Runs the network with the inputVector and returns the outputVector

---

saveToFile

public void saveToFile(java.lang.String fileName)
                throws java.io.IOException

Saves the neural network

---

save

public void save(java.io.BufferedWriter output)
          throws java.io.IOException

Saves the neural network

---

toString

public java.lang.String toString()

retrurns a String representation of the net

---

inputSize

public int inputSize()

returns the number of inputs of the net without counting the bias

---

outputSize

public int outputSize()

returns the number of outputs of the net without counting the bias

---

getNextStringTokenizer

private static java.util.StringTokenizer getNextStringTokenizer(java.io.BufferedReader input,
                                                                java.lang.String token)
                                                         throws java.io.IOException,
                                                                java.util.NoSuchElementException

---

createWeights

private static float[][] createWeights(java.io.BufferedReader input,
                                       int layerSize,
                                       int previousLayerSize)
                                throws java.io.IOException,
                                       java.util.NoSuchElementException

---

netToIntArrey

public int[] netToIntArrey()

returns a int[] which represents a netework the weightValues are converted to int with the Float.floatToIntBits() method.

---

addIntToArray

private static int[] addIntToArray(int[] array,
                                   int element,
                                   int pos)

---

Overview  Package   Class  Use  Deprecated  Index
Home >> All >> AI >> [ NeuralNetworks overview ]	PREV CLASS  NEXT CLASS
SUMMARY:  JAVADOC |  SOURCE |  DOWNLOAD | NESTED | FIELD | CONSTR | METHOD	DETAIL: FIELD | CONSTR | METHOD

---