Package elki.evaluation.clustering.internal

Class ConcordantPairsGammaTau

java.lang.Object

elki.evaluation.clustering.internal.ConcordantPairsGammaTau

All Implemented Interfaces:

Evaluator, ResultProcessor

@Reference(authors="F. B. Baker, L. J. Hubert",
           title="Measuring the Power of Hierarchical Cluster Analysis",
           booktitle="Journal of the American Statistical Association, 70(349)",
           url="https://doi.org/10.1080/01621459.1975.10480256",
           bibkey="doi:10.1080/01621459.1975.10480256")
public class ConcordantPairsGammaTau
extends java.lang.Object
implements Evaluator

Compute the Gamma Criterion of a data set.

References:

F. B. Baker, L. J. Hubert
Measuring the Power of Hierarchical Cluster Analysis
Journal of the American Statistical Association, 70(349)

Tau measures:

F. J. Rohlf
Methods of comparing classifications
Annual Review of Ecology and Systematics

The runtime complexity of this measure is O(n*n*log(n)).

Since:

0.7.0

Author:

Stephan Baier, Erich Schubert

Nested Class Summary

Nested Classes

Modifier and Type	Class	Description
static class	ConcordantPairsGammaTau.Par	Parameterization class.

Field Summary

Fields

Modifier and Type	Field	Description
private PrimitiveDistance<? super NumberVector>	distance	Distance function to use.
private java.lang.String	key	Key for logging statistics.
private static Logging	LOG	Logger for debug output.
private NoiseHandling	noiseHandling	Option for noise handling.

Constructor Summary

Constructors

Constructor	Description
ConcordantPairsGammaTau(PrimitiveDistance<? super NumberVector> distance, NoiseHandling noiseHandling)	Constructor.

Method Summary

Modifier and Type	Method	Description
double	computeTau(long c, long d, double m, long wd, long bd)	Compute the Tau correlation measure
protected double[]	computeWithinDistances(Relation<? extends NumberVector> rel, java.util.List<? extends Cluster<?>> clusters, int withinPairs)
protected int	countTies(double[] withinDistances, int[] withinTies)	Count (and annotate) the number of tied values.
double	evaluateClustering(Relation<? extends NumberVector> rel, Clustering<?> c)	Evaluate a single clustering.
void	processNewResult(java.lang.Object result)	Process a result.

Methods inherited from class java.lang.Object

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

Field Detail

LOG

private static final Logging LOG

Logger for debug output.

noiseHandling

private NoiseHandling noiseHandling

Option for noise handling.

distance

private PrimitiveDistance<? super NumberVector> distance

Distance function to use.

key

private java.lang.String key

Key for logging statistics.

Constructor Detail

ConcordantPairsGammaTau

public ConcordantPairsGammaTau(PrimitiveDistance<? super NumberVector> distance,
                               NoiseHandling noiseHandling)

Constructor.

Parameters:

distance - Distance function

noiseHandling - Control noise handling

Method Detail

evaluateClustering

public double evaluateClustering(Relation<? extends NumberVector> rel,
                                 Clustering<?> c)

Evaluate a single clustering.

Parameters:

rel - Data relation

c - Clustering

Returns:

Gamma index

countTies

protected int countTies(double[] withinDistances,
                        int[] withinTies)

Count (and annotate) the number of tied values.

Parameters:

withinDistances - Distances array

withinTies - Output array of tie counts.

Returns:

Number of tied values.

computeWithinDistances

protected double[] computeWithinDistances(Relation<? extends NumberVector> rel,
                                          java.util.List<? extends Cluster<?>> clusters,
                                          int withinPairs)

computeTau

@Reference(authors="F. J. Rohlf",
           title="Methods of comparing classifications",
           booktitle="Annual Review of Ecology and Systematics",
           url="https://doi.org/10.1146/annurev.es.05.110174.000533",
           bibkey="doi:10.1146/annurev.es.05.110174.000533")
public double computeTau(long c,
                         long d,
                         double m,
                         long wd,
                         long bd)

Compute the Tau correlation measure

Parameters:

c - Concordant pairs

d - Discordant pairs

m - Total number of pairs

wd - Number of within distances

bd - Number of between distances

Returns:

Gamma plus statistic

processNewResult

public void processNewResult(java.lang.Object result)

Description copied from interface: ResultProcessor

Process a result.

Specified by:

processNewResult in interface ResultProcessor

Parameters:

result - Newly added result subtree.