Class GreedyClustering<T>

java.lang.Object

org.ojalgo.data.cluster.GreedyClustering<T>

All Implemented Interfaces:

ClusteringAlgorithm<T>

public final class GreedyClustering<T>
extends Object
implements ClusteringAlgorithm<T>

Greedy clustering algorithm. Assigns each item to the nearest centroid, creating new centroids as needed. Will only pass through the data once. The centroids are recalculated as the clusters are updated (not with every single update, but continuously during the process).

Field Summary

Fields

Modifier and Type	Field	Description
private final List<T>	myCentroids
private final Function<Collection<T>, T>	myCentroidUpdater
private final ToDoubleBiFunction<T,T>	myDistanceCalculator
private final double	myDistanceThreshold
private final List<AtomicInteger>	myUpdates

Constructor Summary

Constructors

Constructor	Description
GreedyClustering(Function<Collection<T>, T> centroidUpdater, ToDoubleBiFunction<T,T> distanceCalculator, double distanceThreshold)

Method Summary

Modifier and Type	Method	Description
List<Set<T>>	cluster(Collection<T> input)	Partitions the given items into clusters.
(package private) List<T>	getCentroids()

Methods inherited from class Object

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

Field Details

myCentroids

private final List<T> myCentroids

myUpdates

private final List<AtomicInteger> myUpdates

myCentroidUpdater

private final Function<Collection<T>, T> myCentroidUpdater

myDistanceCalculator

private final ToDoubleBiFunction<T,T> myDistanceCalculator

myDistanceThreshold

private final double myDistanceThreshold

Constructor Details

GreedyClustering

public GreedyClustering(Function<Collection<T>, T> centroidUpdater,
 ToDoubleBiFunction<T,T> distanceCalculator,
 double distanceThreshold)

Parameters:

centroidUpdater - The update function should return a new centroid based on a collection of points (the set of items in a cluster).

distanceCalculator - A function that calculates the distance between two points.

distanceThreshold - The maximum distance between a point and a centroid for the point to be assigned to that cluster. The points are always assigned to the cluster of the nearest centroid among the already existing clusters. This threshold determines when a new cluster should be created.

Method Details

cluster

public List<Set<T>> cluster(Collection<T> input)

Description copied from interface: ClusteringAlgorithm

Partitions the given items into clusters.

Specified by:

cluster in interface ClusteringAlgorithm<T>

Parameters:

input - The items to cluster; must not be null. May be empty.

Returns:

A list of clusters; each element of the list is a Set<T> representing one cluster and containing its members. Sets are non-empty and pairwise disjoint; ordering is unspecified.

getCentroids

List<T> getCentroids()