neural.net
Class BackPropagator

java.lang.Object
  neural.net.BackPropagator

All Implemented Interfaces:

java.io.Serializable

public class BackPropagator
extends java.lang.Object
implements java.io.Serializable

This class performs back-propagation on a Network to update/train the weights of the Neurons within each Layer in the Network

Version:

1.0.0

Author:

Jonathan Reimels

See Also:

Serialized Form

Constructor Summary

BackPropagator(Network network, double learningRate)
Constructor

Method Summary

void	addInputOutput(double[] inputs, double[] expectedOutputs) Add an input and the expected output for this input
java.util.ArrayList<java.lang.Double>	getAverageErrors() The averages from each input within the last run
double	runAndUpdate() Run the input through an iteration, and update the weights of each neuron after the run of each input

Methods inherited from class java.lang.Object

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

Constructor Detail

BackPropagator

public BackPropagator(Network network,
                      double learningRate)

Constructor

Parameters:

network - - The Network to operate on

learningRate - - The learning rate to use (should be between 0 and 1)

Method Detail

getAverageErrors

public java.util.ArrayList<java.lang.Double> getAverageErrors()

The averages from each input within the last run

Returns:

averages

addInputOutput

public void addInputOutput(double[] inputs,
                           double[] expectedOutputs)
                    throws java.lang.Exception

Add an input and the expected output for this input

Parameters:

inputs - - input to use

expectedOutputs - - output expected from this input

Throws:

java.lang.Exception - - throws an exception if the input size doesn't match that expected by the Network, or the output size doesn't match that returned by the Network

runAndUpdate

public double runAndUpdate()
                    throws java.lang.Exception

Run the input through an iteration, and update the weights of each neuron after the run of each input

Returns:

average error across running all inputs

Throws:

java.lang.Exception