org.encog.neural.flat.train.prop
Class TrainFlatNetworkSCG

java.lang.Object
  org.encog.neural.flat.train.prop.TrainFlatNetworkProp
      org.encog.neural.flat.train.prop.TrainFlatNetworkSCG

All Implemented Interfaces:

TrainFlatNetwork

public class TrainFlatNetworkSCG
extends TrainFlatNetworkProp

Train a network using scaled conjugate gradient.

Field Summary

protected static double	FIRST_LAMBDA The starting value for lambda.
protected static double	FIRST_SIGMA The starting value for sigma.

Fields inherited from class org.encog.neural.flat.train.prop.TrainFlatNetworkProp

currentError, gradients, lastError, network

Constructor Summary

TrainFlatNetworkSCG(FlatNetwork network, MLDataSet training)
Construct the training object.

Method Summary

void	calculateGradients() Calculate the gradients.
void	initOthers() Perform training method specific init.
void	iteration() Perform one iteration.
double	updateWeight(double[] gradients, double[] lastGradient, int index) Update the weights.

Methods inherited from class org.encog.neural.flat.train.prop.TrainFlatNetworkProp

finishTraining, fixFlatSpot, getError, getErrorFunction, getIteration, getLastGradient, getNetwork, getNumThreads, getTraining, iteration, learn, learnLimited, report, setErrorFunction, setIteration, setNumThreads

Methods inherited from class java.lang.Object

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

Field Detail

FIRST_SIGMA

protected static final double FIRST_SIGMA

The starting value for sigma.

See Also:

Constant Field Values

FIRST_LAMBDA

protected static final double FIRST_LAMBDA

The starting value for lambda.

See Also:

Constant Field Values

Constructor Detail

TrainFlatNetworkSCG

public TrainFlatNetworkSCG(FlatNetwork network,
                           MLDataSet training)

Construct the training object.

Parameters:

network - The network to train.

training - The training data to use.

Method Detail

calculateGradients

public void calculateGradients()

Calculate the gradients. They are normalized as well.

Overrides:

calculateGradients in class TrainFlatNetworkProp

iteration

public void iteration()

Perform one iteration.

Specified by:

iteration in interface TrainFlatNetwork

Overrides:

iteration in class TrainFlatNetworkProp

updateWeight

public double updateWeight(double[] gradients,
                           double[] lastGradient,
                           int index)

Update the weights.

Specified by:

updateWeight in class TrainFlatNetworkProp

Parameters:

gradients - The current gradients.

lastGradient - The last gradients.

index - The weight index being updated.

Returns:

The new weight value.

initOthers

public void initOthers()

Perform training method specific init.

Specified by:

initOthers in class TrainFlatNetworkProp