org.cognity.neural.net

Class AbstractFeedforwardNet

java.lang.Object

org.cognity.neural.net.AbstractNeuralNet

org.cognity.neural.net.AbstractFeedforwardNet

All Implemented Interfaces:

Serializable, FeedforwardNet, NeuralNet

Direct Known Subclasses:

MlpNet

public abstract class AbstractFeedforwardNet
extends AbstractNeuralNet
implements FeedforwardNet

This class implements a feedforward neural network. It implements a basic computional logic for such networks, so subclasses will usually implement only the builder object.

Author:

Artur Markiewicz

See Also:

Serialized Form

Constructor Summary

Constructors

Modifier	Constructor and Description
protected	AbstractFeedforwardNet(Layer inputLayer, Layer outputLayer, Set<Layer> layers, Set<Link> links, List<Layer> orderedLayers)

Method Summary

Methods

Modifier and Type	Method and Description
double[]	compute(double[] input) Performs computations for this network.
List<Layer>	getOrderedLayers() Returns a list of all layers in the network.

Methods inherited from class org.cognity.neural.net.AbstractNeuralNet

getInputLayer, getInputLinks, getInputSize, getLayerCount, getLayers, getLinks, getOneInputLink, getOneOutputLink, getOutputLayer, getOutputLinks, getOutputSize, hasInputLinks, hasOutputLinks, initialize, toString

Methods inherited from class java.lang.Object

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

Methods inherited from interface org.cognity.neural.net.NeuralNet

getInputLayer, getInputLinks, getInputSize, getLayerCount, getLayers, getLinks, getOneInputLink, getOneOutputLink, getOutputLayer, getOutputLinks, getOutputSize, hasInputLinks, hasOutputLinks

Constructor Detail

AbstractFeedforwardNet

protected AbstractFeedforwardNet(Layer inputLayer,
                      Layer outputLayer,
                      Set<Layer> layers,
                      Set<Link> links,
                      List<Layer> orderedLayers)

Method Detail

compute

public double[] compute(double[] input)

Description copied from interface: NeuralNet

Performs computations for this network.

Specified by:

compute in interface NeuralNet

Parameters:

input - The input vector. Cannot be null. The size of the vector must be equal to the size of the input layer.

Returns:

the output vector

getOrderedLayers

public List<Layer> getOrderedLayers()

Description copied from interface: FeedforwardNet

Returns a list of all layers in the network. The order of layers in the list corresponds to the topological order of layers. I.e. layers are sorted from the input layer to the output layer.

Specified by:

getOrderedLayers in interface FeedforwardNet

Returns:

a list of all layers in the network