jsat.clustering.hierarchical

Class DivisiveGlobalClusterer

java.lang.Object

jsat.clustering.ClustererBase

jsat.clustering.KClustererBase

jsat.clustering.hierarchical.DivisiveGlobalClusterer

All Implemented Interfaces:

java.io.Serializable, Clusterer, KClusterer

public class DivisiveGlobalClusterer
extends KClustererBase

DivisiveGlobalClusterer is a hierarchical clustering method that works by splitting the data set into sub trees from the top down. Unlike many top-up methods, such as SimpleHAC, top-down methods require another clustering method to perform the splitting at each iteration. If the base method is not deterministic, then the top-down method will not be deterministic.
Like many HAC methods, DivisiveGlobalClusterer will store the merge order of the clusters so that the clustering results for many k can be obtained. It is limited to the range of clusters successfully computed before.

Specifically, DivisiveGlobalClusterer greedily chooses the cluster to split based on an evaluation of all resulting clusters after a split. Because of this global search of the world, DivisiveLocalClusterer has can make a good estimate of the number of clusters in the data set. The quality of this result is dependent on the accuracy of the ClusterEvaluation used. This quality comes at the cost of execution speed, as more and more large evaluations of the whole dataset are needed at each iteration. If execution speed is more important, DivisiveLocalClusterer should be used instead, which requires only a fixed number of evaluations per iteration.

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor and Description
DivisiveGlobalClusterer(DivisiveGlobalClusterer toCopy) Copy constructor
DivisiveGlobalClusterer(KClusterer baseClusterer, ClusterEvaluation clusterEvaluation)

Method Summary

Modifier and Type	Method and Description
DivisiveGlobalClusterer	clone()
int[]	cluster(DataSet dataSet, boolean parallel, int[] designations) Performs clustering on the given data set.
int[]	cluster(DataSet dataSet, int clusters, boolean parallel, int[] designations)
int[]	cluster(DataSet dataSet, int lowK, int highK, boolean parallel, int[] designations)
int[]	clusterSplit(int targetK) Returns the clustering results for a specific k number of clusters for a previously computed data set.

Methods inherited from class jsat.clustering.ClustererBase

createClusterListFromAssignmentArray, getDatapointsFromCluster

Methods inherited from class java.lang.Object

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

Methods inherited from interface jsat.clustering.KClusterer

cluster, cluster, cluster, cluster, cluster, cluster

Methods inherited from interface jsat.clustering.Clusterer

cluster, cluster, cluster, supportsWeightedData

Constructor Detail

DivisiveGlobalClusterer

public DivisiveGlobalClusterer(KClusterer baseClusterer,
                               ClusterEvaluation clusterEvaluation)

DivisiveGlobalClusterer

public DivisiveGlobalClusterer(DivisiveGlobalClusterer toCopy)

Copy constructor

Parameters:

toCopy - the object to copy

Method Detail

cluster

public int[] cluster(DataSet dataSet,
                     boolean parallel,
                     int[] designations)

Description copied from interface: Clusterer

Performs clustering on the given data set. Parameters may be estimated by the method, or other heuristics performed.

Parameters:

dataSet - the data set to perform clustering on

parallel - true if multiple threads should be used to perform clustering. false if it should be done in a single threaded manner.

designations - the array which will contain the designated values. The array will be altered and returned by the function. If null is given, a new array will be created and returned.

Returns:

the int[]

cluster

public int[] cluster(DataSet dataSet,
                     int clusters,
                     boolean parallel,
                     int[] designations)

cluster

public int[] cluster(DataSet dataSet,
                     int lowK,
                     int highK,
                     boolean parallel,
                     int[] designations)

clusterSplit

public int[] clusterSplit(int targetK)

Returns the clustering results for a specific k number of clusters for a previously computed data set. If the data set did not compute up to the value k null will be returned.

Parameters:

targetK - the number of clusters to get the result for.

Returns:

an array containing the assignments for each cluster in the original data set.

Throws:

ClusterFailureException - if no prior data set had been clustered

clone

public DivisiveGlobalClusterer clone()

Specified by:

clone in interface Clusterer

Specified by:

clone in interface KClusterer

Specified by:

clone in class KClustererBase