jsat.clustering

Class ClustererBase

java.lang.Object

jsat.clustering.ClustererBase

All Implemented Interfaces:

Serializable, Clusterer

Direct Known Subclasses:

DBSCAN, FLAME, HDBSCAN, KClustererBase, LSDBC, MeanShift, OPTICS

public abstract class ClustererBase
extends Object
implements Clusterer

A base foundation that provides an implementation of cluster(jsat.DataSet) and cluster(jsat.DataSet, java.util.concurrent.ExecutorService) using their int array counterparts.

By default it is assumed that a cluster does not support weighted data. If this is incorrect, you need to overwrite the supportsWeightedData() method.

Author:

Edward Raff

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor and Description
ClustererBase()

Method Summary

Modifier and Type	Method and Description
abstract Clusterer	clone()
List<List<DataPoint>>	cluster(DataSet dataSet) Performs clustering on the given data set.
List<List<DataPoint>>	cluster(DataSet dataSet, ExecutorService threadpool) Performs clustering on the given data set.
static List<List<DataPoint>>	createClusterListFromAssignmentArray(int[] assignments, DataSet dataSet) Convenient helper method.
static List<DataPoint>	getDatapointsFromCluster(int c, int[] assignments, DataSet dataSet, int[] indexFrom) Gets a list of the datapoints in a data set that belong to the indicated cluster
boolean	supportsWeightedData() Indicates whether the model knows how to cluster using weighted data points.

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

Constructor Detail

ClustererBase

public ClustererBase()

Method Detail

cluster

public List<List<DataPoint>> cluster(DataSet dataSet)

Description copied from interface: Clusterer

Performs clustering on the given data set. Parameters may be estimated by the method, or other heuristics performed.

Specified by:

cluster in interface Clusterer

Parameters:

dataSet - the data set to perform clustering on

Returns:

A list of clusters found by this method.

cluster

public List<List<DataPoint>> cluster(DataSet dataSet,
                                     ExecutorService threadpool)

Description copied from interface: Clusterer

Performs clustering on the given data set. Parameters may be estimated by the method, or other heuristics performed.

Specified by:

cluster in interface Clusterer

Parameters:

dataSet - the data set to perform clustering on

threadpool - a source of threads to run tasks

Returns:

list of clusters found by this method.

createClusterListFromAssignmentArray

public static List<List<DataPoint>> createClusterListFromAssignmentArray(int[] assignments,
                                                                         DataSet dataSet)

Convenient helper method. A list of lists to represent a cluster may be desirable. In such a case, this method will take in an array of cluster assignments, and return a list of lists.

Parameters:

assignments - the array containing cluster assignments

dataSet - the original data set, with data in the same order as was used to create the assignments array

Returns:

a List of lists where each list contains the data points for one cluster, and the lists are in order by cluster id.

getDatapointsFromCluster

public static List<DataPoint> getDatapointsFromCluster(int c,
                                                       int[] assignments,
                                                       DataSet dataSet,
                                                       int[] indexFrom)

Gets a list of the datapoints in a data set that belong to the indicated cluster

Parameters:

c - the cluster ID to get the datapoints for

assignments - the array containing cluster assignments

dataSet - the data set to get the points from

indexFrom - stores the index from the original dataset that the datapoint is from, such that the item at index i in the returned list can be found in the original dataset at index indexFrom[i]. May be null

Returns:

a list of datapoints that were assignment to the designated cluster

supportsWeightedData

public boolean supportsWeightedData()

Description copied from interface: Clusterer

Indicates whether the model knows how to cluster using weighted data points. If it does, the model will train assuming the weights. The values returned by this method may change depending on the parameters set for the model.

Specified by:

supportsWeightedData in interface Clusterer

Returns:

true if the model supports weighted data, false otherwise

clone

public abstract Clusterer clone()

Specified by:

clone in interface Clusterer

Overrides:

clone in class Object