|
|||||||||
| Home >> All >> AI >> [ NeuralNetworks overview ] | PREV CLASS NEXT CLASS | ||||||||
SUMMARY: JAVADOC | SOURCE | DOWNLOAD | NESTED | FIELD | CONSTR | METHOD |
DETAIL: FIELD | CONSTR | METHOD | ||||||||
AI.NeuralNetworks
Class BackErrorPropagation

java.lang.ObjectAI.NeuralNetworks.FeedForwardNetwork
AI.NeuralNetworks.BackErrorPropagation
- All Implemented Interfaces:
- NeuralNetworkTeacher
- Direct Known Subclasses:
- DistributedBackErrorPropagation
- public class BackErrorPropagation
- extends FeedForwardNetwork
- implements NeuralNetworkTeacher
- extends FeedForwardNetwork
This Class is a nerual network with a multiple layer feed forward architecture where the back error propagation could be run
| Field Summary | |
protected static java.lang.String |
COMMENT_TOKEN
|
protected static float |
DEFAULT_LEARNING_RATE
|
protected static long |
DEFAULT_MAXIMUM_EPOCH
|
protected static float |
DEFAULT_MOMENTUM_RATE
|
protected static float |
DEFAULT_STOP_ERROR
|
protected static java.lang.String |
END_TOKEN
|
protected static java.lang.String |
INPUT_TOKEN
|
protected float |
learningRate
|
protected long |
maximumEpoch
|
protected float |
momentumRate
|
protected Notify |
notifyError
|
protected static java.lang.String |
OUTPUT_TOKEN
|
protected float |
stopError
|
protected static java.lang.String |
VECTOR_TOKEN
|
| Fields inherited from class AI.NeuralNetworks.FeedForwardNetwork |
HEXADECIMAL_VERSION, INPUT_LAYER_TOKEN, layers, NUMBER_OF_LAYERS_TOKEN, VERSION, VERSION_TOKEN, WEIGHTED_LAYER_TOKEN, WEIGHTS_TOKEN |
| Constructor Summary | |
protected |
BackErrorPropagation()
this construcctor should not be used it is here for inheriterence only |
|
BackErrorPropagation(FeedForwardNetwork network)
Creates a new BackErrorPropagation instance |
|
BackErrorPropagation(FeedForwardNetwork network,
float learningRate,
float momentumRate,
long maximumEpoch,
float stopError)
Creates a new BackErrorPropagation instance |
| Method Summary | |
void |
addDeltaWeights()
Adds the deltaWeights to the weights |
void |
adjustDeltaWeights(float[] input,
float[] expectedOutput)
Calculates the delta weights for this pattern and adds it to the acumlated deltas |
private static java.util.StringTokenizer |
getNextStringTokenizer(java.io.BufferedReader input,
java.lang.String token)
|
static java.util.LinkedList |
getVectors(java.io.BufferedReader input)
|
static java.util.LinkedList |
getVectorsFormFile(java.lang.String fileName)
|
static int[] |
parseStructure(java.lang.String stringStructure)
converts a string composed by tokens that could be converted to int to an array of int |
void |
resetDeltaWeights()
Sets the acumulated delta weights to 0 |
void |
setLearnningRate(float newRate)
Sets the learnning rate |
void |
setMaximumEpoch(int newMaxEpoch)
Sets the maximum epoch |
void |
setNotifyError(Notify notifyError)
Sets an implementation of Notify, that will be notified of the error of each epoch |
void |
setStopError(float newStopError)
Sets the stop error |
float |
trainNetwork(java.util.LinkedList trainingPatterns)
Trains the network with the given patterns |
| Methods inherited from class AI.NeuralNetworks.FeedForwardNetwork |
inputSize, netToIntArrey, outputSize, runVector, save, saveToFile, toString |
| Methods inherited from class java.lang.Object |
clone, equals, finalize, getClass, hashCode, notify, notifyAll, wait, wait, wait |
| Field Detail |
COMMENT_TOKEN
protected static final java.lang.String COMMENT_TOKEN
- See Also:
- Constant Field Values
INPUT_TOKEN
protected static final java.lang.String INPUT_TOKEN
- See Also:
- Constant Field Values
OUTPUT_TOKEN
protected static final java.lang.String OUTPUT_TOKEN
- See Also:
- Constant Field Values
VECTOR_TOKEN
protected static final java.lang.String VECTOR_TOKEN
- See Also:
- Constant Field Values
END_TOKEN
protected static final java.lang.String END_TOKEN
- See Also:
- Constant Field Values
DEFAULT_LEARNING_RATE
protected static float DEFAULT_LEARNING_RATE
DEFAULT_MOMENTUM_RATE
protected static float DEFAULT_MOMENTUM_RATE
DEFAULT_MAXIMUM_EPOCH
protected static long DEFAULT_MAXIMUM_EPOCH
DEFAULT_STOP_ERROR
protected static float DEFAULT_STOP_ERROR
learningRate
protected float learningRate
momentumRate
protected float momentumRate
maximumEpoch
protected long maximumEpoch
stopError
protected float stopError
notifyError
protected Notify notifyError
| Constructor Detail |
BackErrorPropagation
public BackErrorPropagation(FeedForwardNetwork network, float learningRate, float momentumRate, long maximumEpoch, float stopError)
- Creates a new BackErrorPropagation instance
BackErrorPropagation
public BackErrorPropagation(FeedForwardNetwork network)
- Creates a new BackErrorPropagation instance
BackErrorPropagation
protected BackErrorPropagation()
- this construcctor should not be used it is here for inheriterence only
| Method Detail |
trainNetwork
public float trainNetwork(java.util.LinkedList trainingPatterns) throws UnexpectedInputArraySizeException
- Trains the network with the given patterns
- Specified by:
trainNetworkin interfaceNeuralNetworkTeacher
resetDeltaWeights
public void resetDeltaWeights()
- Sets the acumulated delta weights to 0
adjustDeltaWeights
public void adjustDeltaWeights(float[] input,
float[] expectedOutput)
throws UnexpectedInputArraySizeException
- Calculates the delta weights for this pattern and adds it to the acumlated deltas
addDeltaWeights
public void addDeltaWeights()
- Adds the deltaWeights to the weights
setNotifyError
public void setNotifyError(Notify notifyError)
- Sets an implementation of Notify, that will be notified of the error of each epoch
setLearnningRate
public void setLearnningRate(float newRate)
- Sets the learnning rate
setMaximumEpoch
public void setMaximumEpoch(int newMaxEpoch)
- Sets the maximum epoch
setStopError
public void setStopError(float newStopError)
- Sets the stop error
getVectorsFormFile
public static java.util.LinkedList getVectorsFormFile(java.lang.String fileName)
getVectors
public static java.util.LinkedList getVectors(java.io.BufferedReader input)
parseStructure
public static int[] parseStructure(java.lang.String stringStructure)
- converts a string composed by tokens that could be converted to int to an array of int
getNextStringTokenizer
private static java.util.StringTokenizer getNextStringTokenizer(java.io.BufferedReader input, java.lang.String token) throws java.io.IOException, java.util.NoSuchElementException
|
|||||||||
| Home >> All >> AI >> [ NeuralNetworks overview ] | PREV CLASS NEXT CLASS | ||||||||
SUMMARY: JAVADOC | SOURCE | DOWNLOAD | NESTED | FIELD | CONSTR | METHOD |
DETAIL: FIELD | CONSTR | METHOD | ||||||||
JAVADOC