org.encog.neural.networks.training.propagation
Class CalculatePartialDerivative

java.lang.Object
  org.encog.neural.networks.training.propagation.CalculatePartialDerivative

public class CalculatePartialDerivative
extends java.lang.Object

Class that is used to calculate the partial derivatives for the error for individual layers of a neural network. This calculation must be performed by each of the propagation techniques.

Author:

jheaton

Constructor Summary

CalculatePartialDerivative()

Method Summary

void	calculateError(NeuralOutputHolder output, PropagationLevel fromLevel, PropagationLevel toLevel) Calculate the partial derivative of the error for a layer.

Methods inherited from class java.lang.Object

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

Constructor Detail

CalculatePartialDerivative

public CalculatePartialDerivative()

Method Detail

calculateError

public void calculateError(NeuralOutputHolder output,
                           PropagationLevel fromLevel,
                           PropagationLevel toLevel)

Calculate the partial derivative of the error for a layer.

Parameters:

output - A holder that contains the output from all of the layers.

fromLevel - The source level.

toLevel - The target level.