jsat.clustering.kmeans

Class MiniBatchKMeans

java.lang.Object

jsat.clustering.ClustererBase

jsat.clustering.KClustererBase

jsat.clustering.kmeans.MiniBatchKMeans

All Implemented Interfaces:

java.io.Serializable, Clusterer, KClusterer

public class MiniBatchKMeans
extends KClustererBase

Implements the mini-batch algorithms for k-means. This is a stochastic algorithm, so it does not find the global solution. This implementation is parallel, but only the methods that specify the exact number of clusters are supported.

See: Sculley, D. (2010). Web-scale k-means clustering. Proceedings of the 19th international conference on World wide web (pp. 1177–1178). New York, New York, USA: ACM Press. doi:10.1145/1772690.1772862

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor and Description
MiniBatchKMeans(DistanceMetric dm, int batchSize, int iterations) Creates a new Mini-Batch k-Means object that uses k-means++ for seed selection.
MiniBatchKMeans(DistanceMetric dm, int batchSize, int iterations, SeedSelectionMethods.SeedSelection seedSelection) Creates a new Mini-Batch k-Means object
MiniBatchKMeans(int batchSize, int iterations) Creates a new Mini-Batch k-Means object that uses k-means++ for seed selection and uses the EuclideanDistance.
MiniBatchKMeans(MiniBatchKMeans toCopy) Copy constructor

Method Summary

Modifier and Type	Method and Description
MiniBatchKMeans	clone()
int[]	cluster(DataSet dataSet, boolean parallel, int[] designations) Performs clustering on the given data set.
int[]	cluster(DataSet dataSet, 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	getBatchSize() Returns the batch size used at each iteration
DistanceMetric	getDistanceMetric() Returns the distance metric used for determining the nearest cluster center
int	getIterations() Returns the number of mini-batch iterations used
java.util.List<Vec>	getMeans() Returns the raw list of means that were used for each class.
SeedSelectionMethods.SeedSelection	getSeedSelection() Returns the method of seed selection to use
void	setBatchSize(int batchSize) Sets the batch size to use at each iteration.
void	setDistanceMetric(DistanceMetric dm) Sets the distance metric used for determining the nearest cluster center
void	setIterations(int iterations) Sets the number of mini-batch iterations to perform
void	setSeedSelection(SeedSelectionMethods.SeedSelection seedSelection) Sets the method of selecting the initial data points to seed the clustering algorithm.
void	setStoreMeans(boolean storeMeans) If set to true the computed means will be stored after clustering is completed, and can then be retrieved using getMeans().

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, supportsWeightedData

Constructor Detail

MiniBatchKMeans

public MiniBatchKMeans(int batchSize,
                       int iterations)

Creates a new Mini-Batch k-Means object that uses k-means++ for seed selection and uses the EuclideanDistance.

Parameters:

batchSize - the mini-batch size

iterations - the number of mini batches to perform

MiniBatchKMeans

public MiniBatchKMeans(DistanceMetric dm,
                       int batchSize,
                       int iterations)

Creates a new Mini-Batch k-Means object that uses k-means++ for seed selection.

Parameters:

dm - the distance metric to use

batchSize - the mini-batch size

iterations - the number of mini batches to perform

MiniBatchKMeans

public MiniBatchKMeans(DistanceMetric dm,
                       int batchSize,
                       int iterations,
                       SeedSelectionMethods.SeedSelection seedSelection)

Creates a new Mini-Batch k-Means object

Parameters:

dm - the distance metric to use

batchSize - the mini-batch size

iterations - the number of mini batches to perform

seedSelection - the seed selection algorithm to initiate clustering

MiniBatchKMeans

public MiniBatchKMeans(MiniBatchKMeans toCopy)

Copy constructor

Parameters:

toCopy - the object to copy

Method Detail

setStoreMeans

public void setStoreMeans(boolean storeMeans)

If set to true the computed means will be stored after clustering is completed, and can then be retrieved using getMeans().

Parameters:

storeMeans - true if the means should be stored for later, false to discard them once clustering is complete.

getMeans

public java.util.List<Vec> getMeans()

Returns the raw list of means that were used for each class.

Returns:

the list of means for each class

setDistanceMetric

public void setDistanceMetric(DistanceMetric dm)

Sets the distance metric used for determining the nearest cluster center

Parameters:

dm - the distance metric to use

getDistanceMetric

public DistanceMetric getDistanceMetric()

Returns the distance metric used for determining the nearest cluster center

Returns:

the distance metric in use

setBatchSize

public void setBatchSize(int batchSize)

Sets the batch size to use at each iteration. Increasing the batch size can improve the resulting clustering, but increases computational cost at each iteration.
If the batch size is set equal to or larger than data set size, it reduces to the naive k-means algorithm.

Parameters:

batchSize - the number of points to use at each iteration

getBatchSize

public int getBatchSize()

Returns the batch size used at each iteration

Returns:

the batch size in use

setIterations

public void setIterations(int iterations)

Sets the number of mini-batch iterations to perform

Parameters:

iterations - the number of algorithm iterations to perform

getIterations

public int getIterations()

Returns the number of mini-batch iterations used

Returns:

the number of algorithm iterations that will be used

setSeedSelection

public void setSeedSelection(SeedSelectionMethods.SeedSelection seedSelection)

Sets the method of selecting the initial data points to seed the clustering algorithm.

Parameters:

seedSelection - the seed selection algorithm to use

getSeedSelection

public SeedSelectionMethods.SeedSelection getSeedSelection()

Returns the method of seed selection to use

Returns:

the method of seed selection to use

cluster

public int[] cluster(DataSet dataSet,
                     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

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:

an array indicating for each value indicating the cluster designation. This is the same array as designations, or a new one if the input array was null

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)

clone

public MiniBatchKMeans clone()

Specified by:

clone in interface Clusterer

Specified by:

clone in interface KClusterer

Specified by:

clone in class KClustererBase