cc.glsn.v15.neuralnet

Class RecurrentBrain

java.lang.Object

cc.glsn.v15.neuralnet.Brain

cc.glsn.v15.neuralnet.RecurrentBrain

All Implemented Interfaces:

Serializable

public class RecurrentBrain
extends Brain

Extends Brain to become an feedback neural network rather than a simple feed forward. The specified numbers of feeback links are added as outputs and those outputs are feed back in as inputs.

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor and Description
RecurrentBrain(NetFunction functg, LinkedList<Integer> layerSizes, int feedbackLinks)
RecurrentBrain(NetFunction functg, LinkedList<Integer> layerSizes, String inputMappingStyle, int feedbackLinks)

Method Summary

Methods

Modifier and Type	Method and Description
void	backPropogate(double alpha, Map<String,Double> targets) Do back propogation on a group of output targets at once.
void	backPropogate(String name, double alpha, double target) Do back propogation learning with the given alpha and target.
void	resetFeeds(double val)
void	saveFeeds() Optional.
void	updateFeeds() Copies all output feedback values to the inputs Should be run before a new set of inputs are set.

Methods inherited from class cc.glsn.v15.neuralnet.Brain

addInput, addInputGroup, addOutput, getAllOutputs, getInputNode, getInternalLinkCount, getOutput, incSerial, removeOutput, setAllInput, setHiddenLayers, setInput, toString

Methods inherited from class java.lang.Object

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

Constructor Detail

RecurrentBrain

public RecurrentBrain(NetFunction functg,
              LinkedList<Integer> layerSizes,
              int feedbackLinks)

RecurrentBrain

public RecurrentBrain(NetFunction functg,
              LinkedList<Integer> layerSizes,
              String inputMappingStyle,
              int feedbackLinks)

Method Detail

resetFeeds

public void resetFeeds(double val)

saveFeeds

public void saveFeeds()

Optional. Saves the feedback line values for later update. This allows for intermediate processes like back propogation to change things while still keeping the right feed values for the next go.

updateFeeds

public void updateFeeds()

Copies all output feedback values to the inputs Should be run before a new set of inputs are set. Order should be: (optional) saveFeeds() backPropogate() updateFeeds() setInput()s getOutput()s

backPropogate

public void backPropogate(double alpha,
                 Map<String,Double> targets)

Description copied from class: Brain

Do back propogation on a group of output targets at once. Time complexity n where n is number of neuron links

Overrides:

backPropogate in class Brain

Parameters:

alpha - - learning factor. Should be positive and less than one. 0.01 is a good starting point.

targets - a map of string(output names) to doubles (target output values) to learn on

backPropogate

public void backPropogate(String name,
                 double alpha,
                 double target)

Description copied from class: Brain

Do back propogation learning with the given alpha and target. Time complexity n where n is number of neuron links

Overrides:

backPropogate in class Brain

Parameters:

name - - Output name to learn

alpha - - The learning factor. Should be positive and less than one. 0.01 is a good starting point.

target - - What the output value should be for the set inputs