org.expasy.mzjava.core.ms.cluster

Class KMeansPlusPlusClusterBuilder<V>

java.lang.Object

org.expasy.mzjava.core.ms.cluster.KMeansPlusPlusClusterBuilder<V>

All Implemented Interfaces:

ClusterBuilder<V>

public class KMeansPlusPlusClusterBuilder<V>
extends Object
implements ClusterBuilder<V>

KMeans cluster builder that uses the algorithm provided by commons math KMeansPlusPlusClusterer.

Version:

sqrt -1

Author:

Oliver Horlacher

Constructor Summary

Constructors

Constructor and Description
KMeansPlusPlusClusterBuilder(int numClusters)
KMeansPlusPlusClusterBuilder(int numClusters, com.google.common.base.Optional<Long> seed)

Method Summary

Methods

Modifier and Type	Method and Description
Collection<Set<V>>	cluster(SimilarityGraph<V> graph) Cluster the vertices given a SimilarityGraph
Collection<Set<V>>	cluster(SimilarityGraph<V> graph, Collection<Set<V>> startingClusters) Cluster the vertices in graph using the startingClusters as a starting point.

Methods inherited from class java.lang.Object

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

Constructor Detail

KMeansPlusPlusClusterBuilder

public KMeansPlusPlusClusterBuilder(int numClusters)

KMeansPlusPlusClusterBuilder

public KMeansPlusPlusClusterBuilder(int numClusters,
                            com.google.common.base.Optional<Long> seed)

Method Detail

cluster

public Collection<Set<V>> cluster(SimilarityGraph<V> graph)

Description copied from interface: ClusterBuilder

Cluster the vertices given a SimilarityGraph

Specified by:

cluster in interface ClusterBuilder<V>

Parameters:

graph - the graph

Returns:

collection containing sets of vertices that belong to the same cluster

cluster

public Collection<Set<V>> cluster(SimilarityGraph<V> graph,
                         Collection<Set<V>> startingClusters)

Description copied from interface: ClusterBuilder

Cluster the vertices in graph using the startingClusters as a starting point.

Cluster builders that where the starting clusters have no impact on clustering, such as the threshold cluster builder, can delegate to the cluster(SimilarityGraph graph) method.

Specified by:

cluster in interface ClusterBuilder<V>

Parameters:

graph - the graph

startingClusters - the clusters to start with

Returns:

collection containing sets of vertices that belong to the same cluster