org.encog.neural.networks.layers
Class FeedforwardLayer

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

All Implemented Interfaces:

java.io.Serializable, Layer, EncogPersistedObject

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

FeedforwardLayer: This class represents one layer in a feed forward neural network. This layer could be input, output, or hidden, depending on its placement inside of the FeedforwardNetwork class. An activation function can also be specified. Usually all layers in a neural network will use the same activation function. By default this class uses the sigmoid activation function.

See Also:

Serialized Form

Constructor Summary

FeedforwardLayer(ActivationFunction thresholdFunction, int neuronCount)
Construct this layer with a non-default threshold function.

FeedforwardLayer(int neuronCount)
Construct this layer with a sigmoid threshold function.

Method Summary

FeedforwardLayer	cloneStructure() Clone the structure of this layer, but do not copy any matrix data.
NeuralData	compute(NeuralData pattern) Compute the outputs for this layer given the input pattern.
ActivationFunction	getActivationFunction()
void	prune(int neuron) Prune one of the neurons from this layer.
void	setMatrix(Matrix matrix) Assign a new weight and threshold matrix to this layer.
void	setNext(Layer next) Set the next layer.
java.lang.String	toString()

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

getFire, getFire, getMatrix, getMatrixSize, getNeuronCount, getNext, getPrevious, hasMatrix, isHidden, isInput, isOutput, reset, setFire, setFire, setPrevious

Methods inherited from class java.lang.Object

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

Constructor Detail

FeedforwardLayer

public FeedforwardLayer(ActivationFunction thresholdFunction,
                        int neuronCount)

Construct this layer with a non-default threshold function.

Parameters:

thresholdFunction - The threshold function to use.

neuronCount - How many neurons in this layer.

FeedforwardLayer

public FeedforwardLayer(int neuronCount)

Construct this layer with a sigmoid threshold function.

Parameters:

neuronCount - How many neurons in this layer.

Method Detail

cloneStructure

public FeedforwardLayer cloneStructure()

Clone the structure of this layer, but do not copy any matrix data.

Returns:

The cloned layer.

compute

public NeuralData compute(NeuralData pattern)

Compute the outputs for this layer given the input pattern. The output is also stored in the fire instance variable.

Specified by:

compute in interface Layer

Overrides:

compute in class BasicLayer

Parameters:

pattern - The input pattern.

Returns:

The output from this layer.

getActivationFunction

public ActivationFunction getActivationFunction()

Returns:

The activation function for this layer.

prune

public void prune(int neuron)

Prune one of the neurons from this layer. Remove all entries in this weight matrix and other layers.

Parameters:

neuron - The neuron to prune. Zero specifies the first neuron.

setMatrix

public void setMatrix(Matrix matrix)

Assign a new weight and threshold matrix to this layer.

Specified by:

setMatrix in interface Layer

Overrides:

setMatrix in class BasicLayer

Parameters:

matrix - The new matrix.

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.

toString

public java.lang.String toString()

Overrides:

toString in class java.lang.Object

Returns:

The string form of the layer.