Package org.aksw.limes.core.evaluation.qualititativeMeasures

Class APRF

java.lang.Object

org.aksw.limes.core.evaluation.qualititativeMeasures.APRF

All Implemented Interfaces:

IQualitativeMeasure

Direct Known Subclasses:

Accuracy, APseudoPRF, AUC, FMeasure, Precision, Recall

public abstract class APRF
extends Object
implements IQualitativeMeasure

This class is an abstract class for the Precision, Recall and F-Measure classes.
It contains set of methods that calculate the values of true-positive, false-positive, true-negative and false-negative which are used by evaluators classes to evaluate the mappings results.

Since:

1.0

Version:

1.0

Author:

Klaus Lyko (lyko@informatik.uni-leipzig.de), Mofeed Hassan (mounir@informatik.uni-leipzig.de)

Constructor Summary

Constructors

Constructor	Description
APRF()

Method Summary

Modifier and Type	Method	Description
abstract double	calculate(AMapping predictions, GoldStandard goldStandard)	The Abstract method to be implemented for calculating the accuracy of the machine learning predictions compared to a gold standard
static double	falseNegative(AMapping predictions, AMapping goldStandard)	The method calculates the false-negative results such that the result is claimed by a machine learning as a negative one and the claim is false.
static double	trueFalsePositive(AMapping predictions, AMapping goldStandard, boolean truePositive)	The method calculates either the true positive or the false positive results which are defined as true-positive: the results classified as positive and the classification is correct while false-positive: the results classified as positive and the classification is incorrect.
static double	trueNegative(AMapping predictions, GoldStandard goldStandard)

Methods inherited from class java.lang.Object

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

Constructor Detail

APRF

public APRF()

Method Detail

trueFalsePositive

public static double trueFalsePositive(AMapping predictions,
                                       AMapping goldStandard,
                                       boolean truePositive)

The method calculates either the true positive or the false positive results which are defined as true-positive: the results classified as positive and the classification is correct while false-positive: the results classified as positive and the classification is incorrect.

Parameters:

predictions - The predictions provided by a machine learning algorithm

goldStandard - It contains the gold standard (reference mapping) combined with the source and target URIs

truePositive - A flag switches the calculation between true positive (=true) and false positive(=false)

Returns:

double - This returns either True positive or False positive based on the flag value

falseNegative

public static double falseNegative(AMapping predictions,
                                   AMapping goldStandard)

The method calculates the false-negative results such that the result is claimed by a machine learning as a negative one and the claim is false.

Parameters:

predictions - The predictions provided by a machine learning algorithm

goldStandard - It contains the gold standard (reference mapping) combined with the source and target URIs

Returns:

double This returns the number of false negative links

trueNegative

public static double trueNegative(AMapping predictions,
                                  GoldStandard goldStandard)

calculate

public abstract double calculate(AMapping predictions,
                                 GoldStandard goldStandard)

The Abstract method to be implemented for calculating the accuracy of the machine learning predictions compared to a gold standard

Specified by:

calculate in interface IQualitativeMeasure

Parameters:

predictions - The predictions provided by a machine learning algorithm

goldStandard - It contains the gold standard (reference mapping) combined with the source and target URIs

Returns:

double - This returns the calculated accuracy