de.lmu.ifi.dbs.elki.evaluation.clustering

Class ClusterContingencyTable

java.lang.Object

de.lmu.ifi.dbs.elki.evaluation.clustering.ClusterContingencyTable

public class ClusterContingencyTable
extends java.lang.Object

Class storing the contingency table and related data on two clusterings.

Since:

0.5.0

Author:

Erich Schubert

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	ClusterContingencyTable.Util Utility class.

Field Summary

Fields

Modifier and Type	Field and Description
protected BCubed	bcubed BCubed measures
protected boolean	breakNoiseClusters Noise cluster handling
protected int[][]	contingency Contingency matrix
protected EditDistance	edit Edit-Distance measures
protected Entropy	entropy Entropy-based measures
protected long[]	noise1 Noise flags
protected long[]	noise2 Noise flags
protected PairCounting	paircount Pair counting measures
protected boolean	selfPairing Self pairing
protected int	size1 Number of clusters.
protected int	size2 Number of clusters.
protected SetMatchingPurity	smp Set matching purity measures

Constructor Summary

Constructors

Constructor and Description
ClusterContingencyTable(boolean selfPairing, boolean breakNoiseClusters) Constructor.

Method Summary

Modifier and Type	Method and Description
MeanVariance	averageSymmetricGini() Compute the average Gini for each cluster (in both clusterings - symmetric).
BCubed	getBCubed() The BCubed based measures
EditDistance	getEdit() Get (compute) the edit-distance based measures
Entropy	getEntropy() Get (compute) the entropy based measures
PairCounting	getPaircount() Get (compute) the pair counting measures.
SetMatchingPurity	getSetMatching() The set-matching measures
void	process(Clustering<?> result1, Clustering<?> result2) Process two clustering results.
java.lang.String	toString()

Methods inherited from class java.lang.Object

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

Field Detail

breakNoiseClusters

protected boolean breakNoiseClusters

Noise cluster handling

selfPairing

protected boolean selfPairing

Self pairing

size1

protected int size1

Number of clusters.

size2

protected int size2

Number of clusters.

contingency

protected int[][] contingency

Contingency matrix

noise1

protected long[] noise1

Noise flags

noise2

protected long[] noise2

Noise flags

paircount

protected PairCounting paircount

Pair counting measures

entropy

protected Entropy entropy

Entropy-based measures

smp

protected SetMatchingPurity smp

Set matching purity measures

edit

protected EditDistance edit

Edit-Distance measures

bcubed

protected BCubed bcubed

BCubed measures

Constructor Detail

ClusterContingencyTable

public ClusterContingencyTable(boolean selfPairing,
                               boolean breakNoiseClusters)

Constructor.

Parameters:

selfPairing - Build self-pairs

breakNoiseClusters - Break noise clusters into individual objects

Method Detail

process

public void process(Clustering<?> result1,
                    Clustering<?> result2)

Process two clustering results.

Parameters:

result1 - First clustering

result2 - Second clustering

toString

public java.lang.String toString()

Overrides:

toString in class java.lang.Object

getPaircount

public PairCounting getPaircount()

Get (compute) the pair counting measures.

Returns:

Pair counting measures

getEntropy

public Entropy getEntropy()

Get (compute) the entropy based measures

Returns:

Entropy based measures

getEdit

public EditDistance getEdit()

Get (compute) the edit-distance based measures

Returns:

Edit-distance based measures

getBCubed

public BCubed getBCubed()

The BCubed based measures

Returns:

BCubed measures

getSetMatching

public SetMatchingPurity getSetMatching()

The set-matching measures

Returns:

Set-Matching measures

averageSymmetricGini

public MeanVariance averageSymmetricGini()

Compute the average Gini for each cluster (in both clusterings - symmetric).

Returns:

Mean and variance of Gini