Package elki.evaluation.clustering

Class EvaluateClustering

java.lang.Object

elki.evaluation.clustering.EvaluateClustering

All Implemented Interfaces:

Evaluator, ResultProcessor

public class EvaluateClustering
extends java.lang.Object
implements Evaluator

Evaluate a clustering result by comparing it to an existing cluster label.

Since:

0.4.0

Author:

Erich Schubert

Nested Class Summary

Nested Classes

Modifier and Type	Class	Description
static class	EvaluateClustering.Par	Parameterization class.
static class	EvaluateClustering.ScoreResult	Result object for outlier score judgements.

Field Summary

Fields

Modifier and Type	Field	Description
private static Logging	LOG	Logger for debug output.
private boolean	noiseSpecialHandling	Apply special handling to noise "clusters".
private ClusteringAlgorithm<?>	referencealg	Reference algorithm.
private boolean	selfPairing	Use self-pairing in pair-counting measures

Constructor Summary

Constructors

Constructor	Description
EvaluateClustering(ClusteringAlgorithm<?> referencealg, boolean noiseSpecialHandling, boolean selfPairing)	Constructor.

Method Summary

Modifier and Type	Method	Description
static double	evaluateRanking(ScoreEvaluation eval, Cluster<?> clus, DoubleDBIDList ranking)	Evaluate given a cluster (of positive elements) and a scoring list.
protected void	evaluteResult(Database db, Clustering<?> c, Clustering<?> refc)	Evaluate a clustering result.
private boolean	isReferenceResult(Clustering<?> t)	Test if a clustering result is a valid reference result.
void	processNewResult(java.lang.Object newResult)	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.

referencealg

private ClusteringAlgorithm<?> referencealg

Reference algorithm.

noiseSpecialHandling

private boolean noiseSpecialHandling

Apply special handling to noise "clusters".

selfPairing

private boolean selfPairing

Use self-pairing in pair-counting measures

Constructor Detail

EvaluateClustering

public EvaluateClustering(ClusteringAlgorithm<?> referencealg,
                          boolean noiseSpecialHandling,
                          boolean selfPairing)

Constructor.

Parameters:

referencealg - Reference clustering

noiseSpecialHandling - Noise handling flag

selfPairing - Self-pairing flag

Method Detail

evaluateRanking

public static double evaluateRanking(ScoreEvaluation eval,
                                     Cluster<?> clus,
                                     DoubleDBIDList ranking)

Evaluate given a cluster (of positive elements) and a scoring list.

Parameters:

eval - Evaluation method

clus - Cluster object

ranking - Object ranking

Returns:

Score

processNewResult

public void processNewResult(java.lang.Object newResult)

Description copied from interface: ResultProcessor

Process a result.

Specified by:

processNewResult in interface ResultProcessor

Parameters:

newResult - Newly added result subtree.

evaluteResult

protected void evaluteResult(Database db,
                             Clustering<?> c,
                             Clustering<?> refc)

Evaluate a clustering result.

Parameters:

db - Database

c - Clustering

refc - Reference clustering

isReferenceResult

private boolean isReferenceResult(Clustering<?> t)

Test if a clustering result is a valid reference result.

Parameters:

t - Clustering to test.

Returns:

true if it is considered to be a reference result.