Package schule.ngb.zm.ml

Class NeuronLayer

java.lang.Object

schule.ngb.zm.ml.NeuronLayer

All Implemented Interfaces:

Function<MLMatrix,MLMatrix>

public class NeuronLayer
extends Object
implements Function<MLMatrix,MLMatrix>

Implementierung einer Neuronenebene in einem Neuonalen Netz.

Eine Ebene besteht aus einer Anzahl an Neuronen die jeweils eine Anzahl Eingänge haben. Die Eingänge erhalten als Signal die Ausgabe der vorherigen Ebene und berechnen die Ausgabe des jeweiligen Neurons.

Constructor Summary

Constructors

Constructor	Description
NeuronLayer(int neurons, int inputs)

Method Summary

Modifier and Type	Method	Description
<V> Function<MLMatrix,V>	andThen(Function<? super MLMatrix,? extends V> after)
MLMatrix	apply(MLMatrix inputs)
void	backprop(MLMatrix expected, double learningRate)
<V> Function<V,MLMatrix>	compose(Function<? super V,? extends MLMatrix> before)
void	connect(NeuronLayer prev, NeuronLayer next)
static NeuronLayer	fromArray(double[][] weights, boolean transpose)
static NeuronLayer	fromArray(double[][] weights, double[] biases, boolean transpose)
MLMatrix	getBiases()
int	getInputCount()
MLMatrix	getLastOutput()
int	getNeuronCount()
NeuronLayer	getNextLayer()
NeuronLayer	getPreviousLayer()
MLMatrix	getWeights()
boolean	hasNextLayer()
boolean	hasPreviousLayer()
void	setNextLayer(NeuronLayer pNextLayer)
void	setPreviousLayer(NeuronLayer pPreviousLayer)
void	setWeights(MLMatrix newWeights)
String	toString()

Methods inherited from class java.lang.Object

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

Constructor Details

NeuronLayer

public NeuronLayer(int neurons,
 int inputs)

Method Details

fromArray

public static NeuronLayer fromArray(double[][] weights,
 boolean transpose)

fromArray

public static NeuronLayer fromArray(double[][] weights,
 double[] biases,
 boolean transpose)

connect

public void connect(NeuronLayer prev,
 NeuronLayer next)

getPreviousLayer

public NeuronLayer getPreviousLayer()

hasPreviousLayer

public boolean hasPreviousLayer()

setPreviousLayer

public void setPreviousLayer(NeuronLayer pPreviousLayer)

getNextLayer

public NeuronLayer getNextLayer()

hasNextLayer

public boolean hasNextLayer()

setNextLayer

public void setNextLayer(NeuronLayer pNextLayer)

getWeights

public MLMatrix getWeights()

getBiases

public MLMatrix getBiases()

getNeuronCount

public int getNeuronCount()

getInputCount

public int getInputCount()

getLastOutput

public MLMatrix getLastOutput()

setWeights

public void setWeights(MLMatrix newWeights)

toString

public String toString()

Overrides:

toString in class Object

apply

public MLMatrix apply(MLMatrix inputs)

Specified by:

apply in interface Function<MLMatrix,MLMatrix>

compose

public <V> Function<V,MLMatrix> compose(Function<? super V,? extends MLMatrix> before)

Specified by:

compose in interface Function<MLMatrix,MLMatrix>

andThen

public <V> Function<MLMatrix,V> andThen(Function<? super MLMatrix,? extends V> after)

Specified by:

andThen in interface Function<MLMatrix,MLMatrix>

backprop

public void backprop(MLMatrix expected,
 double learningRate)