mulan.evaluation.measure
Class AveragePrecision

java.lang.Object
  mulan.evaluation.measure.MeasureBase
      mulan.evaluation.measure.RankingMeasureBase
          mulan.evaluation.measure.AveragePrecision

All Implemented Interfaces:

Serializable, Measure

public class AveragePrecision
extends RankingMeasureBase

Implementation of the average precision measure. It evaluates the average fraction of labels ranked above a particular relevant label, that are actually relevant.

Version:

2010.11.05

Author:

Jozef Vilcek, Grigorios Tsoumakas

See Also:

Serialized Form

Field Summary

Fields inherited from class mulan.evaluation.measure.RankingMeasureBase

count, sum

Constructor Summary

AveragePrecision()

Method Summary

double	getIdealValue() Gets an 'ideal' value of a measure.
String	getName() Gets the name of a measure.
protected void	updateRanking(int[] ranking, boolean[] trueLabels) Updates the measure based on an example

Methods inherited from class mulan.evaluation.measure.RankingMeasureBase

getValue, reset, updateInternal

Methods inherited from class mulan.evaluation.measure.MeasureBase

makeCopy, toString, update

Methods inherited from class java.lang.Object

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

Constructor Detail

AveragePrecision

public AveragePrecision()

Method Detail

getName

public String getName()

Description copied from interface: Measure

Gets the name of a measure.

Returns:

the name of a measure.

getIdealValue

public double getIdealValue()

Description copied from interface: Measure

Gets an 'ideal' value of a measure. The 'ideal' means, that the value represents the best achievable performance of a learner for an prediction of a multi-label task and associated true labels.

Returns:

the ideal value

updateRanking

protected void updateRanking(int[] ranking,
                             boolean[] trueLabels)

Description copied from class: RankingMeasureBase

Updates the measure based on an example

Specified by:

updateRanking in class RankingMeasureBase

Parameters:

ranking - the predicted ranking

trueLabels - the ground truth