edu.ucla.sspace.clustering

Class Assignments

java.lang.Object

edu.ucla.sspace.clustering.Assignments

All Implemented Interfaces:

Iterable<Assignment>

public class Assignments
extends Object
implements Iterable<Assignment>

The return value for all Clustering implementations. This class records the number of clusters created, the assignments for each value, and helper methods for constructing the centroids of a cluster.

Author:

Keith Stevens

Constructor Summary

Constructors

Constructor and Description
Assignments(int numClusters, Assignment[] initialAssignments) Creates a new Assignments instance that takes ownership of the initialAssignments array.
Assignments(int numClusters, Assignment[] initialAssignments, Matrix matrix) Creates a new Assignments instance that takes ownership of the initialAssignments array.
Assignments(int numClusters, int numAssignments) Creates a new Assignments instance that can hold up to numAssignments Assignments.
Assignments(int numClusters, int numAssignments, Matrix matrix) Creates a new Assignments instance that can hold up to numAssignments Assignments.

Method Summary

Methods

Modifier and Type	Method and Description
Assignment[]	assignments() Returns the array of Assignment objects.
List<Set<Integer>>	clusters() Returns the data point indices assigned to each cluster.
int[]	clusterSizes()
Assignment	get(int i) Returns the Assignment object at index i.
DoubleVector[]	getCentroids() Returns an array of dense centroid vectors of each discovered cluster which are scaled according the the number of data points asisgned to that cluster.
SparseDoubleVector[]	getSparseCentroids() Returns an array of sparse centroid vectors of each discovered cluster which are scaled according the the number of data points asisgned to that cluster.
Iterator<Assignment>	iterator() Returns an iterator over the Assignment objects stored.
int	numClusters() Returns the number of clusters.
void	set(int i, Assignment assignment) Sets Assignment i to have value assignment.
int	size() Returns the number of Assignment objects stored.

Methods inherited from class java.lang.Object

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

Constructor Detail

Assignments

public Assignments(int numClusters,
           int numAssignments)

Creates a new Assignments instance that can hold up to numAssignments Assignments. This assumes that the data matrix will not be accessible. Calls to getCentroids() will fail when using this constructor.

Assignments

public Assignments(int numClusters,
           int numAssignments,
           Matrix matrix)

Creates a new Assignments instance that can hold up to numAssignments Assignments.

Assignments

public Assignments(int numClusters,
           Assignment[] initialAssignments)

Creates a new Assignments instance that takes ownership of the initialAssignments array. This assumes that the data matrix will not be accessible. Calls to getCentroids() will fail when using this constructor.

Assignments

public Assignments(int numClusters,
           Assignment[] initialAssignments,
           Matrix matrix)

Creates a new Assignments instance that takes ownership of the initialAssignments array.

Method Detail

set

public void set(int i,
       Assignment assignment)

Sets Assignment i to have value assignment.

size

public int size()

Returns the number of Assignment objects stored.

iterator

public Iterator<Assignment> iterator()

Returns an iterator over the Assignment objects stored.

Specified by:

iterator in interface Iterable<Assignment>

get

public Assignment get(int i)

Returns the Assignment object at index i.

numClusters

public int numClusters()

Returns the number of clusters.

assignments

public Assignment[] assignments()

Returns the array of Assignment objects.

clusters

public List<Set<Integer>> clusters()

Returns the data point indices assigned to each cluster.

getCentroids

public DoubleVector[] getCentroids()

Returns an array of dense centroid vectors of each discovered cluster which are scaled according the the number of data points asisgned to that cluster. Note that this method assumes that the original Matrix holding the data points contains rows of feature vectors.

getSparseCentroids

public SparseDoubleVector[] getSparseCentroids()

Returns an array of sparse centroid vectors of each discovered cluster which are scaled according the the number of data points asisgned to that cluster. This assumes that the original Matrix is sparse. Note that this method assumes that the original Matrix holding the data points contains rows of feature vectors.

clusterSizes

public int[] clusterSizes()