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

Class SampleKMeansInitialization<V extends NumberVector>

java.lang.Object

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

de.lmu.ifi.dbs.elki.algorithm.clustering.kmeans.initialization.SampleKMeansInitialization<V>

Type Parameters:

V - Vector type

All Implemented Interfaces:

KMeansInitialization

@Alias(value="de.lmu.ifi.dbs.elki.algorithm.clustering.kmeans.SampleKMeansInitialization")
 @Reference(authors="P. S. Bradley, U. M. Fayyad",
           title="Refining Initial Points for K-Means Clustering",
           booktitle="Proc. 15th Int. Conf. on Machine Learning (ICML 1998)",
           bibkey="DBLP:conf/icml/BradleyF98")
public class SampleKMeansInitialization<V extends NumberVector>
extends AbstractKMeansInitialization

Initialize k-means by running k-means on a sample of the data set only.

Reference:

The idea of finding centers on a sample can be found in:

P. S. Bradley, U. M. Fayyad
Refining Initial Points for K-Means Clustering
Proc. 15th Int. Conf. on Machine Learning (ICML 1998)

But Bradley and Fayyad also suggest to repeat this multiple times. This implementation uses a single attempt only.

Since:

0.6.0

Author:

Erich Schubert

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	SampleKMeansInitialization.Parameterizer<V extends NumberVector> Parameterization class.

Field Summary

Fields

Modifier and Type	Field and Description
private KMeans<V,?>	innerkMeans Variant of kMeans to use for initialization.
private double	rate Sample size.

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

rnd

Constructor Summary

Constructors

Constructor and Description
SampleKMeansInitialization(RandomFactory rnd, KMeans<V,?> innerkMeans, double rate) Constructor.

Method Summary

Modifier and Type	Method and Description
double[][]	chooseInitialMeans(Database database, Relation<? extends NumberVector> relation, int k, NumberVectorDistanceFunction<?> distanceFunction) Choose initial means

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

unboxVectors

Methods inherited from class java.lang.Object

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

Field Detail

innerkMeans

private KMeans<V extends NumberVector,?> innerkMeans

Variant of kMeans to use for initialization.

rate

private double rate

Sample size.

Constructor Detail

SampleKMeansInitialization

public SampleKMeansInitialization(RandomFactory rnd,
                                  KMeans<V,?> innerkMeans,
                                  double rate)

Constructor.

Parameters:

rnd - Random generator.

innerkMeans - Inner k-means algorithm.

rate - Sampling rate.

Method Detail

chooseInitialMeans

public double[][] chooseInitialMeans(Database database,
                                     Relation<? extends NumberVector> relation,
                                     int k,
                                     NumberVectorDistanceFunction<?> distanceFunction)

Description copied from interface: KMeansInitialization

Choose initial means

Parameters:

database - Database context

relation - Relation

k - Parameter k

distanceFunction - Distance function

Returns:

List of chosen means for k-means