de.lmu.ifi.dbs.elki.algorithm.clustering.kmeans

Class XMeans<V extends NumberVector,M extends MeanModel>

java.lang.Object

de.lmu.ifi.dbs.elki.algorithm.AbstractAlgorithm<R>

de.lmu.ifi.dbs.elki.algorithm.AbstractNumberVectorDistanceBasedAlgorithm<V,Clustering<M>>

de.lmu.ifi.dbs.elki.algorithm.clustering.kmeans.AbstractKMeans<V,M>

de.lmu.ifi.dbs.elki.algorithm.clustering.kmeans.XMeans<V,M>

Type Parameters:

V - Vector type

M - Model type

All Implemented Interfaces:

Algorithm, ClusteringAlgorithm<Clustering<M>>, KMeans<V,M>, DistanceBasedAlgorithm<V>

@Reference(authors="D. Pelleg, A. Moore",
           booktitle="X-means: Extending K-means with Efficient Estimation on the Number of Clusters",
           title="Proceedings of the 17th International Conference on Machine Learning (ICML 2000)",
           url="http://www.pelleg.org/shared/hp/download/xmeans.ps")
public class XMeans<V extends NumberVector,M extends MeanModel>
extends AbstractKMeans<V,M>

X-means: Extending K-means with Efficient Estimation on the Number of Clusters. Note: this implementation does currently not use a k-d-tree for acceleration. Also note that k_max is not a hard threshold - the algorithm can return up to 2*k_max clusters! Reference:

D. Pelleg, A. Moore:
X-means: Extending K-means with Efficient Estimation on the Number of Clusters
In: Proceedings of the 17th International Conference on Machine Learning (ICML 2000)

Author:

Tibor Goldschwendt, Erich Schubert

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	XMeans.Parameterizer<V extends NumberVector,M extends MeanModel> Parameterization class.

Field Summary

Fields

Modifier and Type	Field and Description
(package private) KMeansQualityMeasure<V>	informationCriterion Information criterion to choose the better split.
private KMeans<V,M>	innerKMeans Inner k-means algorithm.
private int	k Effective number of clusters, minimum and maximum.
private int	k_max Effective number of clusters, minimum and maximum.
private int	k_min Effective number of clusters, minimum and maximum.
private static String	KEY Key for statistics logging.
private static Logging	LOG The logger for this class.
(package private) RandomFactory	rnd Random factory.
(package private) PredefinedInitialMeans	splitInitializer Initializer for k-means.

Fields inherited from class de.lmu.ifi.dbs.elki.algorithm.clustering.kmeans.AbstractKMeans

initializer, maxiter

Fields inherited from class de.lmu.ifi.dbs.elki.algorithm.AbstractNumberVectorDistanceBasedAlgorithm

distanceFunction

Fields inherited from interface de.lmu.ifi.dbs.elki.algorithm.clustering.kmeans.KMeans

INIT_ID, K_ID, MAXITER_ID, SEED_ID

Fields inherited from interface de.lmu.ifi.dbs.elki.algorithm.DistanceBasedAlgorithm

DISTANCE_FUNCTION_ID

Constructor Summary

Constructors

Constructor and Description
XMeans(NumberVectorDistanceFunction<? super V> distanceFunction, int k_min, int k_max, int maxiter, KMeans<V,M> innerKMeans, KMeansInitialization<? super V> initializer, PredefinedInitialMeans splitInitializer, KMeansQualityMeasure<V> informationCriterion, RandomFactory random) Constructor.

Method Summary

Methods

Modifier and Type	Method and Description
TypeInformation[]	getInputTypeRestriction() Get the input type restriction used for negotiating the data query.
protected Logging	getLogger() Get the (STATIC) logger for this class.
Clustering<M>	run(Database database, Relation<V> relation) Run the algorithm on a database and relation.
protected List<? extends NumberVector>	splitCentroid(Cluster<? extends MeanModel> parentCluster, Relation<V> relation) Split an existing centroid into two initial centers.
protected List<Cluster<M>>	splitCluster(Cluster<M> parentCluster, Database database, Relation<V> relation) Conditionally splits the clusters based on the information criterion.

Methods inherited from class de.lmu.ifi.dbs.elki.algorithm.clustering.kmeans.AbstractKMeans

assignToNearestCluster, incrementalUpdateMean, logVarstat, macQueenIterate, means, medians, setDistanceFunction, setK

Methods inherited from class de.lmu.ifi.dbs.elki.algorithm.AbstractNumberVectorDistanceBasedAlgorithm

getDistanceFunction

Methods inherited from class de.lmu.ifi.dbs.elki.algorithm.AbstractAlgorithm

makeParameterDistanceFunction, run

Methods inherited from class java.lang.Object

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

Methods inherited from interface de.lmu.ifi.dbs.elki.algorithm.clustering.ClusteringAlgorithm

run

Methods inherited from interface de.lmu.ifi.dbs.elki.algorithm.DistanceBasedAlgorithm

getDistanceFunction

Field Detail

LOG

private static final Logging LOG

The logger for this class.

KEY

private static final String KEY

Key for statistics logging.

innerKMeans

private KMeans<V extends NumberVector,M extends MeanModel> innerKMeans

Inner k-means algorithm.

k

private int k

Effective number of clusters, minimum and maximum.

k_min

private int k_min

Effective number of clusters, minimum and maximum.

k_max

private int k_max

Effective number of clusters, minimum and maximum.

splitInitializer

PredefinedInitialMeans splitInitializer

Initializer for k-means.

informationCriterion

KMeansQualityMeasure<V extends NumberVector> informationCriterion

Information criterion to choose the better split.

rnd

RandomFactory rnd

Random factory.

Constructor Detail

XMeans

public XMeans(NumberVectorDistanceFunction<? super V> distanceFunction,
      int k_min,
      int k_max,
      int maxiter,
      KMeans<V,M> innerKMeans,
      KMeansInitialization<? super V> initializer,
      PredefinedInitialMeans splitInitializer,
      KMeansQualityMeasure<V> informationCriterion,
      RandomFactory random)

Constructor.

Parameters:

distanceFunction - Distance function

k_min - k_min parameter - minimum number of result clusters

k_max - k_max parameter - maximum number of result clusters

maxiter - Maximum number of iterations each.

innerKMeans - K-Means variant to use inside.

informationCriterion - The information criterion used for the splitting step

random - Random factory

Method Detail

run

public Clustering<M> run(Database database,
                Relation<V> relation)

Run the algorithm on a database and relation.

Parameters:

database - Database to process

relation - Data relation

Returns:

Clustering result.

splitCluster

protected List<Cluster<M>> splitCluster(Cluster<M> parentCluster,
                            Database database,
                            Relation<V> relation)

Conditionally splits the clusters based on the information criterion.

Parameters:

parentCluster - Cluster to split

database - Database

relation - Data relation

Returns:

Parent cluster when split decreases clustering quality or child clusters when split improves clustering.

splitCentroid

protected List<? extends NumberVector> splitCentroid(Cluster<? extends MeanModel> parentCluster,
                                         Relation<V> relation)

Split an existing centroid into two initial centers.

Parameters:

parentCluster - Existing cluster

relation - Data relation

Returns:

List of new centroids

getInputTypeRestriction

public TypeInformation[] getInputTypeRestriction()

Description copied from class: AbstractAlgorithm

Get the input type restriction used for negotiating the data query.

Specified by:

getInputTypeRestriction in interface Algorithm

Overrides:

getInputTypeRestriction in class AbstractKMeans<V extends NumberVector,M extends MeanModel>

Returns:

Type restriction

getLogger

protected Logging getLogger()

Description copied from class: AbstractAlgorithm

Get the (STATIC) logger for this class.

Specified by:

getLogger in class AbstractAlgorithm<Clustering<M extends MeanModel>>

Returns:

the static logger