org.encog.neural.networks.layers
Class SOMLayer

java.lang.Object
  org.encog.neural.networks.layers.BasicLayer
      org.encog.neural.networks.layers.SOMLayer

All Implemented Interfaces:

java.io.Serializable, Layer, EncogPersistedObject

public class SOMLayer
extends BasicLayer
implements java.io.Serializable, EncogPersistedObject

SelfOrganizingMap: The Self Organizing Map, or Kohonen Neural Network, is a special type of neural network that is used to classify input into groups. The SOM makes use of unsupervised training.

See Also:

Serialized Form

Field Summary

static double

VERYSMALL Do not allow patterns to go below this very small number.

Constructor Summary

SOMLayer(int inputCount, NormalizeInput.NormalizationType normalizationType)
The constructor.

Method Summary

NeuralData	compute(NeuralData pattern) Compute the output from this layer.
Persistor	createPersistor() Create a persistor.
NormalizeInput.NormalizationType	getNormalizationType() Get the normalization type.
void	setNeuronCount(int count) Set the neuron count.
void	setNext(Layer next) Set the next layer.
void	setNormalizationType(NormalizeInput.NormalizationType normalizationType) Set the normalization type.

Methods inherited from class org.encog.neural.networks.layers.BasicLayer

getDescription, getFire, getFire, getMatrix, getMatrixSize, getName, getNeuronCount, getNext, getPrevious, hasMatrix, isHidden, isInput, isOutput, reset, setDescription, setFire, setFire, setMatrix, setName, setPrevious

Methods inherited from class java.lang.Object

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

Methods inherited from interface org.encog.neural.persist.EncogPersistedObject

getDescription, getName, setDescription, setName

Field Detail

VERYSMALL

public static final double VERYSMALL

Do not allow patterns to go below this very small number.

See Also:

Constant Field Values

Constructor Detail

SOMLayer

public SOMLayer(int inputCount,
                NormalizeInput.NormalizationType normalizationType)

The constructor.

Parameters:

inputCount - Number of input neurons

normalizationType - The normalization type to use.

Method Detail

compute

public NeuralData compute(NeuralData pattern)

Compute the output from this layer.

Specified by:

compute in interface Layer

Overrides:

compute in class BasicLayer

Parameters:

pattern - The pattern to compute for.

Returns:

The output from the layer.

createPersistor

public Persistor createPersistor()

Create a persistor.

Specified by:

createPersistor in interface EncogPersistedObject

Overrides:

createPersistor in class BasicLayer

Returns:

The newly created persistor.

getNormalizationType

public NormalizeInput.NormalizationType getNormalizationType()

Get the normalization type.

Returns:

The normalization type.

setNeuronCount

public void setNeuronCount(int count)

Set the neuron count.

Overrides:

setNeuronCount in class BasicLayer

Parameters:

count - The neuron count.

setNext

public void setNext(Layer next)

Set the next layer.

Specified by:

setNext in interface Layer

Overrides:

setNext in class BasicLayer

Parameters:

next - the next layer.

setNormalizationType

public void setNormalizationType(NormalizeInput.NormalizationType normalizationType)

Set the normalization type.

Parameters:

normalizationType - the normalizationType to set