org.neuroph.contrib.learning

Class CrossEntropyError

java.lang.Object

org.neuroph.contrib.learning.CrossEntropyError

All Implemented Interfaces:

java.io.Serializable, ErrorFunction

public class CrossEntropyError
extends java.lang.Object
implements ErrorFunction, java.io.Serializable

Special error function which is recommended to be used in classification models

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor and Description
CrossEntropyError()

Method Summary

Modifier and Type	Method and Description
double[]	addPatternError(double[] predictedOutput, double[] targetOutput) Calculates and adds pattern error for the given predicted and target output vector and adds it to total error.
double	getTotalError() Return total network error
void	reset() Sets total error and pattern count to zero.

Methods inherited from class java.lang.Object

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

Constructor Detail

CrossEntropyError

public CrossEntropyError()

Method Detail

getTotalError

public double getTotalError()

Description copied from interface: ErrorFunction

Return total network error

Specified by:

getTotalError in interface ErrorFunction

Returns:

total network error

reset

public void reset()

Description copied from interface: ErrorFunction

Sets total error and pattern count to zero.

Specified by:

reset in interface ErrorFunction

addPatternError

public double[] addPatternError(double[] predictedOutput,
                                double[] targetOutput)

Description copied from interface: ErrorFunction

Calculates and adds pattern error for the given predicted and target output vector and adds it to total error.

Specified by:

addPatternError in interface ErrorFunction

Parameters:

predictedOutput - actual network output

targetOutput - target/desired output

Returns:

returns pattern error vector