com.github.pmerienne.trident.ml.clustering

Class KMeans

java.lang.Object

com.github.pmerienne.trident.ml.clustering.KMeans

All Implemented Interfaces:

Clusterer, Serializable

public class KMeans
extends Object
implements Clusterer, Serializable

Inspired from http://www.cs.princeton.edu/courses/archive/fall08/cos436/Duda/C/sk_means.htm

Author:

pmerienne

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor and Description
KMeans(Integer nbCluster)

Method Summary

Methods

Modifier and Type	Method and Description
Integer	classify(double[] features) Classifies a given sample and updates clusters.
protected double[]	computeDxs() For each features in initFeatures, compute D(x), the distance between x and the nearest center that has already been chosen.
double[]	distribution(double[] features) Predicts the cluster memberships for a given instance.
double[][]	getCentroids() Returns learned clusters as a List of feature's means
protected void	initCentroids() Init clusters using the k-means++ algorithm.
protected void	initIfPossible(double[] features)
protected boolean	isReady()
protected Integer	nearestCentroid(double[] features)
void	reset()
Integer	update(double[] features) Updates clusters with a given sample and return classification.

Methods inherited from class java.lang.Object

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

Constructor Detail

KMeans

public KMeans(Integer nbCluster)

Method Detail

classify

public Integer classify(double[] features)

Description copied from interface: Clusterer

Classifies a given sample and updates clusters.

Specified by:

classify in interface Clusterer

Returns:

update

public Integer update(double[] features)

Description copied from interface: Clusterer

Updates clusters with a given sample and return classification.

Specified by:

update in interface Clusterer

Returns:

distribution

public double[] distribution(double[] features)

Description copied from interface: Clusterer

Predicts the cluster memberships for a given instance.

Specified by:

distribution in interface Clusterer

Returns:

getCentroids

public double[][] getCentroids()

Description copied from interface: Clusterer

Returns learned clusters as a List of feature's means

Specified by:

getCentroids in interface Clusterer

Returns:

nearestCentroid

protected Integer nearestCentroid(double[] features)

isReady

protected boolean isReady()

initIfPossible

protected void initIfPossible(double[] features)

initCentroids

protected void initCentroids()

Init clusters using the k-means++ algorithm. (Arthur, D. and Vassilvitskii, S. (2007). "k-means++: the advantages of careful seeding".

computeDxs

protected double[] computeDxs()

For each features in initFeatures, compute D(x), the distance between x and the nearest center that has already been chosen.

Returns:

reset

public void reset()

Specified by:

reset in interface Clusterer