com.practicalstudies.neural
Class BackpropagationNode

java.lang.Object
  |
  +--com.practicalstudies.neural.Neuron
        |
        +--com.practicalstudies.neural.BackpropagationNode

All Implemented Interfaces:

Node

public class BackpropagationNode

extends Neuron

Represents a neural network node suitable for backpropagation training.

Author:

Michael Wax

See Also:

Neuron

Constructor Summary

BackpropagationNode(Adder adder, ActivationFunction activationFunction, ErrorFunction errorFunction)

Method Summary

double	calculateActivation() Calculate the output of this BackpropagationNode.
double	getError(double actualValue) Returns the error in the output of this BackpropagationNode.

Methods inherited from class com.practicalstudies.neural.Neuron

addInputConnection, addOutputConnection, clearInputConnections, clearOutputConnections, getInputConnection, getInputConnections, getNumberInputConnections, getNumberOutputConnections, getOutput, getOutputConnection, getOutputConnections

Methods inherited from class java.lang.Object

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

Constructor Detail

BackpropagationNode

public BackpropagationNode(Adder adder,
                           ActivationFunction activationFunction,
                           ErrorFunction errorFunction)

Method Detail

calculateActivation

public double calculateActivation()

Calculate the output of this BackpropagationNode.

Overrides:

calculateActivation in class Neuron

Returns:

the output of this neuron

getError

public double getError(double actualValue)

Returns the error in the output of this BackpropagationNode.