edu.udo.cs.yale.operator.performance
Class MinMaxCriterion

java.lang.Object
  edu.udo.cs.yale.operator.AbstractIOObject
      edu.udo.cs.yale.operator.ResultObjectAdapter
          edu.udo.cs.yale.tools.math.Averagable
              edu.udo.cs.yale.operator.performance.PerformanceCriterion
                  edu.udo.cs.yale.operator.performance.MeasuredPerformance
                      edu.udo.cs.yale.operator.performance.MinMaxCriterion

All Implemented Interfaces:

IOObject, ResultObject, java.lang.Cloneable, java.lang.Comparable<Averagable>

public class MinMaxCriterion
extends MeasuredPerformance

This criterion should be used as wrapper around other performance criteria (see MinMaxWrapper). Instead of averages this criterion builds the minimum of the fitness of the criteria which should be averaged. Maximizing the minimum instead of the average could lead to a more robust behaviour and generalization capacity. Furthermore, this criterion can be build arbitrarily weighted linear combinations of both the minimum and the normal average.

Version:

$Id: MinMaxCriterion.java,v 2.12 2006/08/03 14:39:30 ingomierswa Exp $

Author:

Ingo Mierswa

Field Summary

private int	counter
private MeasuredPerformance	delegate
private double	fitness
private double	fitnessSum
private double	value
private double	weight

Constructor Summary

MinMaxCriterion()
Necessary for newInstance() during loading.

MinMaxCriterion(MeasuredPerformance delegate, double weight)

Method Summary

void	buildSingleAverage(Averagable avg) This method should build the average of this and another averagables of the same type.
void	cloneAveragable(Averagable avg) Must be implemented by subclasses such that it copies all values of other to this.
void	countExample(Example example) Counts a single example by invoking the delegates method.
java.lang.String	getDescription() Returns a description of the performance criterion.
int	getExampleCount() Returns the number of data points which was used to determine the criterion value.
double	getFitness() Returns the fitness depending on the value.
double	getMikroAverage() Returns the (current) value of the averagable (the average itself).
double	getMikroVariance() Returns the variance of the averagable.
java.lang.String	getName() Returns the name of this averagable.
void	readCriterionData(java.io.BufferedReader in) Reads the performance data from the given reader.
void	writeCriterionData(java.io.PrintWriter out) Writes the performance criterion into the given writer.

Methods inherited from class edu.udo.cs.yale.operator.performance.MeasuredPerformance

startCounting

Methods inherited from class edu.udo.cs.yale.operator.performance.PerformanceCriterion

compareTo, getMaxFitness, readPerformanceCriterion, writePerformanceCriterion

Methods inherited from class edu.udo.cs.yale.tools.math.Averagable

buildAverage, clone, formatDeviation, formatPercent, formatValue, getAverage, getMakroAverage, getMakroStandardDeviation, getMakroVariance, getMikroStandardDeviation, getStandardDeviation, getVariance, toHTML, toString

Methods inherited from class edu.udo.cs.yale.operator.ResultObjectAdapter

addAction, getActions, getVisualisationComponent, toHTML, toResultString

Methods inherited from class edu.udo.cs.yale.operator.AbstractIOObject

copy, read, read, read, write, write

Methods inherited from class java.lang.Object

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

Methods inherited from interface edu.udo.cs.yale.operator.IOObject

copy, write, write

Field Detail

delegate

private MeasuredPerformance delegate

fitness

private double fitness

fitnessSum

private double fitnessSum

value

private double value

weight

private double weight

counter

private int counter

Constructor Detail

MinMaxCriterion

public MinMaxCriterion()

Necessary for newInstance() during loading.

MinMaxCriterion

public MinMaxCriterion(MeasuredPerformance delegate,
                       double weight)

Method Detail

writeCriterionData

public void writeCriterionData(java.io.PrintWriter out)
                        throws java.io.IOException

Description copied from class: PerformanceCriterion

Writes the performance criterion into the given writer. Please note that the criterion must be written in a human readable format!

Specified by:

writeCriterionData in class PerformanceCriterion

Throws:

java.io.IOException

readCriterionData

public void readCriterionData(java.io.BufferedReader in)
                       throws java.io.IOException

Description copied from class: PerformanceCriterion

Reads the performance data from the given reader.

Specified by:

readCriterionData in class PerformanceCriterion

Throws:

java.io.IOException

getExampleCount

public int getExampleCount()

Description copied from class: PerformanceCriterion

Returns the number of data points which was used to determine the criterion value.

Specified by:

getExampleCount in class PerformanceCriterion

countExample

public void countExample(Example example)

Counts a single example by invoking the delegates method.

Specified by:

countExample in class MeasuredPerformance

getDescription

public java.lang.String getDescription()

Returns a description of the performance criterion.

Specified by:

getDescription in class MeasuredPerformance

getName

public java.lang.String getName()

Description copied from class: Averagable

Returns the name of this averagable.

Specified by:

getName in interface ResultObject

Specified by:

getName in class Averagable

getMikroAverage

public double getMikroAverage()

Description copied from class: Averagable

Returns the (current) value of the averagable (the average itself).

Specified by:

getMikroAverage in class Averagable

getFitness

public double getFitness()

Description copied from class: PerformanceCriterion

Returns the fitness depending on the value. Subclasses should use getAverage() instead of getValue() in this method since usually the makro average should be optmized instead of the mikro average. The mikro average should only be used in the (rare) cases where no makro average is available.

Specified by:

getFitness in class PerformanceCriterion

getMikroVariance

public double getMikroVariance()

Description copied from class: Averagable

Returns the variance of the averagable. The returned value must not be negative. If the averagable does not define a variance this method should return Double.NaN.

Specified by:

getMikroVariance in class Averagable

cloneAveragable

public void cloneAveragable(Averagable avg)

Description copied from class: Averagable

Must be implemented by subclasses such that it copies all values of other to this. When this method is called, it is guaranteed, that other is a subclass of the class of the object it is called on.

Specified by:

cloneAveragable in class Averagable

buildSingleAverage

public void buildSingleAverage(Averagable avg)

Description copied from class: Averagable

This method should build the average of this and another averagables of the same type. The next invocation of getValue() should return the average of this and the given averagable. Please refer to SimpleCriterion for a simple implementation example.

Specified by:

buildSingleAverage in class Averagable