org.encog.neural.networks.training

Class TrainingSetScore

java.lang.Object

org.encog.neural.networks.training.TrainingSetScore

All Implemented Interfaces:

java.io.Serializable, CalculateScore

public class TrainingSetScore
extends java.lang.Object
implements CalculateScore, java.io.Serializable

Calculate a score based on a training set. This class allows simulated annealing or genetic algorithms just as you would any other training set based training method. The method must support regression (MLRegression).

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor and Description
TrainingSetScore(MLDataSet training) Construct a training set score calculation.

Method Summary

Modifier and Type	Method and Description
double	calculateScore(MLMethod method) Calculate the score for the network.
boolean	requireSingleThreaded()
boolean	shouldMinimize() A training set based score should always seek to lower the error, as a result, this method always returns true.

Methods inherited from class java.lang.Object

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

Constructor Detail

TrainingSetScore

public TrainingSetScore(MLDataSet training)

Construct a training set score calculation.

Parameters:

training - The training data to use.

Method Detail

calculateScore

public double calculateScore(MLMethod method)

Calculate the score for the network.

Specified by:

calculateScore in interface CalculateScore

Parameters:

method - The network to calculate for.

Returns:

The score.

shouldMinimize

public boolean shouldMinimize()

A training set based score should always seek to lower the error, as a result, this method always returns true.

Specified by:

shouldMinimize in interface CalculateScore

Returns:

Returns true.

requireSingleThreaded

public boolean requireSingleThreaded()

Specified by:

requireSingleThreaded in interface CalculateScore

Returns:

True, if this score function cannot be done in parallel.