jsat.clustering

Class CLARA

java.lang.Object

jsat.clustering.ClustererBase

jsat.clustering.KClustererBase

jsat.clustering.PAM

jsat.clustering.CLARA

All Implemented Interfaces:

Serializable, Clusterer, KClusterer

public class CLARA
extends PAM

Author:

Edward Raff

See Also:

Serialized Form

Field Summary

Fields inherited from class jsat.clustering.PAM

dm, iterLimit, medoids, rand, repeats, seedSelection, storeMedoids

Constructor Summary

Constructors

Constructor and Description
CLARA()
CLARA(CLARA toCopy) Copy constructor
CLARA(DistanceMetric dm)
CLARA(DistanceMetric dm, Random rand)
CLARA(DistanceMetric dm, Random rand, SeedSelectionMethods.SeedSelection seedSelection)
CLARA(int sampleCount, DistanceMetric dm, Random rand, SeedSelectionMethods.SeedSelection seedSelection)
CLARA(int sampleSize, int sampleCount, DistanceMetric dm, Random rand, SeedSelectionMethods.SeedSelection seedSelection)

Method Summary

Modifier and Type	Method and Description
CLARA	clone()
protected double	cluster(DataSet data, boolean doInit, int[] medioids, int[] assignments, List<Double> cacheAccel) Performs the actual work of PAM.
int[]	cluster(DataSet dataSet, int clusters, int[] designations)
int	getSampleCount()
int	getSampleSize()
void	setSampleCount(int sampleCount) Sets the number of times PAM will be applied to different samples from the data set.
void	setSampleSize(int sampleSize) Sets the number of samples CLARA should take from the data set to perform PAM on.

Methods inherited from class jsat.clustering.PAM

cluster, cluster, cluster, cluster, cluster, getMaxIterations, getMedoids, getSeedSelection, setMaxIterations, setSeedSelection, setStoreMedoids

Methods inherited from class jsat.clustering.KClustererBase

cluster, cluster, cluster, cluster

Methods inherited from class jsat.clustering.ClustererBase

cluster, cluster, createClusterListFromAssignmentArray, getDatapointsFromCluster, supportsWeightedData

Methods inherited from class java.lang.Object

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

Methods inherited from interface jsat.clustering.Clusterer

cluster, cluster, supportsWeightedData

Constructor Detail

CLARA

public CLARA(int sampleSize,
             int sampleCount,
             DistanceMetric dm,
             Random rand,
             SeedSelectionMethods.SeedSelection seedSelection)

CLARA

public CLARA(int sampleCount,
             DistanceMetric dm,
             Random rand,
             SeedSelectionMethods.SeedSelection seedSelection)

CLARA

public CLARA(DistanceMetric dm,
             Random rand,
             SeedSelectionMethods.SeedSelection seedSelection)

CLARA

public CLARA(DistanceMetric dm,
             Random rand)

CLARA

public CLARA(DistanceMetric dm)

CLARA

public CLARA()

CLARA

public CLARA(CLARA toCopy)

Copy constructor

Parameters:

toCopy - the object to copy

Method Detail

getSampleCount

public int getSampleCount()

Returns:

the number of times PAM will be applied to a sample from the data set.

setSampleCount

public void setSampleCount(int sampleCount)

Sets the number of times PAM will be applied to different samples from the data set.

Parameters:

sampleCount - the number of times to apply sampeling.

getSampleSize

public int getSampleSize()

Returns:

the number of samples that will be taken to perform PAM on.

setSampleSize

public void setSampleSize(int sampleSize)

Sets the number of samples CLARA should take from the data set to perform PAM on.

Parameters:

sampleSize - the number of samples to take

cluster

protected double cluster(DataSet data,
                         boolean doInit,
                         int[] medioids,
                         int[] assignments,
                         List<Double> cacheAccel)

Description copied from class: PAM

Performs the actual work of PAM.

Overrides:

cluster in class PAM

Parameters:

data - the data set to apply PAM to

doInit - true if the initialization procedure of training the distance metric, initiating its cache, and selecting he seeds, should be done.

medioids - the array to store the indices that get chosen as the medoids. The length of the array indicates how many medoids should be obtained.

assignments - an array of the same length as data, each value indicating what cluster that point belongs to.

cacheAccel - the pre-computed distance acceleration cache. May be null.

Returns:

the sum of the squared distance from each point to its closest medoids

cluster

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

Specified by:

cluster in interface KClusterer

Overrides:

cluster in class PAM

clone

public CLARA clone()

Specified by:

clone in interface Clusterer

Specified by:

clone in interface KClusterer

Overrides:

clone in class PAM