Package org.apache.commons.math4.legacy.ml.clustering

Class MiniBatchKMeansClusterer<T extends Clusterable>

java.lang.Object

org.apache.commons.math4.legacy.ml.clustering.Clusterer<T>

org.apache.commons.math4.legacy.ml.clustering.KMeansPlusPlusClusterer<T>

org.apache.commons.math4.legacy.ml.clustering.MiniBatchKMeansClusterer<T>

Type Parameters:

T - Type of the points to cluster.

public class MiniBatchKMeansClusterer<T extends Clusterable>
extends KMeansPlusPlusClusterer<T>

Clustering algorithm based on KMeans.

Nested Class Summary

Nested classes/interfaces inherited from class org.apache.commons.math4.legacy.ml.clustering.KMeansPlusPlusClusterer

KMeansPlusPlusClusterer.EmptyClusterStrategy

Constructor Summary

Constructors

Constructor	Description
MiniBatchKMeansClusterer(int k, int maxIterations, int batchSize, int initIterations, int initBatchSize, int maxNoImprovementTimes, DistanceMeasure measure, org.apache.commons.rng.UniformRandomProvider random, KMeansPlusPlusClusterer.EmptyClusterStrategy emptyStrategy)	Build a clusterer.

Method Summary

Modifier and Type	Method	Description
List<CentroidCluster<T>>	cluster(Collection<T> points)	Runs the MiniBatch K-means clustering algorithm.

Methods inherited from class org.apache.commons.math4.legacy.ml.clustering.KMeansPlusPlusClusterer

getMaxIterations, getNumberOfClusters

Methods inherited from class org.apache.commons.math4.legacy.ml.clustering.Clusterer

distance, getDistanceMeasure

Methods inherited from class java.lang.Object

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

Constructor Detail

MiniBatchKMeansClusterer

public MiniBatchKMeansClusterer(int k,
                                int maxIterations,
                                int batchSize,
                                int initIterations,
                                int initBatchSize,
                                int maxNoImprovementTimes,
                                DistanceMeasure measure,
                                org.apache.commons.rng.UniformRandomProvider random,
                                KMeansPlusPlusClusterer.EmptyClusterStrategy emptyStrategy)

Build a clusterer.

Parameters:

k - Number of clusters to split the data into.

maxIterations - Maximum number of iterations to run the algorithm for all the points, The actual number of iterationswill be smaller than maxIterations * size / batchSize, where size is the number of points to cluster. Disabled if negative.

batchSize - Batch size for training iterations.

initIterations - Number of iterations allowed in order to find out the best initial centers.

initBatchSize - Batch size for initializing the clusters centers. A value of 3 * batchSize should be suitable in most cases.

maxNoImprovementTimes - Maximum number of iterations during which no improvement is occuring. A value of 10 is suitable in most cases.

measure - Distance measure.

random - Random generator.

emptyStrategy - Strategy for handling empty clusters that may appear during algorithm iterations.

Method Detail

cluster

public List<CentroidCluster<T>> cluster(Collection<T> points)

Runs the MiniBatch K-means clustering algorithm.

Overrides:

cluster in class KMeansPlusPlusClusterer<T extends Clusterable>

Parameters:

points - Points to cluster (cannot be null).

Returns:

the clusters.

Throws:

MathIllegalArgumentException - if the number of points is smaller than the number of clusters.