org.apache.commons.math3.stat.clustering
Class DBSCANClusterer<T extends Clusterable<T>>

java.lang.Object
  org.apache.commons.math3.stat.clustering.DBSCANClusterer<T>

Type Parameters:

T - type of the points to cluster

public class DBSCANClusterer<T extends Clusterable<T>>
extends Object

DBSCAN (density-based spatial clustering of applications with noise) algorithm.

The DBSCAN algorithm forms clusters based on the idea of density connectivity, i.e. a point p is density connected to another point q, if there exists a chain of points p_i, with i = 1 .. n and p₁ = p and p_n = q, such that each pair <p_i, p_i+1> is directly density-reachable. A point q is directly density-reachable from point p if it is in the ε-neighborhood of this point.

Any point that is not density-reachable from a formed cluster is treated as noise, and will thus not be present in the result.

The algorithm requires two parameters:

• eps: the distance that defines the ε-neighborhood of a point
• minPoints: the minimum number of density-connected points required to form a cluster

Note: as DBSCAN is not a centroid-based clustering algorithm, the resulting Cluster objects will have no defined center, i.e. Cluster.getCenter() will return null.

Since:

3.1

Version:

$Id: DBSCANClusterer.java 1410882 2012-11-18 12:49:49Z tn $

See Also:

DBSCAN (wikipedia), A Density-Based Algorithm for Discovering Clusters in Large Spatial Databases with Noise

Constructor Summary

DBSCANClusterer(double eps, int minPts)
Creates a new instance of a DBSCANClusterer.

Method Summary

List<Cluster<T>>	cluster(Collection<T> points) Performs DBSCAN cluster analysis.
double	getEps() Returns the maximum radius of the neighborhood to be considered.
int	getMinPts() Returns the minimum number of points needed for a cluster.

Methods inherited from class java.lang.Object

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

Constructor Detail

DBSCANClusterer

public DBSCANClusterer(double eps,
                       int minPts)
                throws NotPositiveException

Creates a new instance of a DBSCANClusterer.

Parameters:

eps - maximum radius of the neighborhood to be considered

minPts - minimum number of points needed for a cluster

Throws:

NotPositiveException - if eps < 0.0 or minPts < 0

Method Detail

getEps

public double getEps()

Returns the maximum radius of the neighborhood to be considered.

Returns:

maximum radius of the neighborhood

getMinPts

public int getMinPts()

Returns the minimum number of points needed for a cluster.

Returns:

minimum number of points needed for a cluster

cluster

public List<Cluster<T>> cluster(Collection<T> points)
                                                throws NullArgumentException

Performs DBSCAN cluster analysis.

Note: as DBSCAN is not a centroid-based clustering algorithm, the resulting Cluster objects will have no defined center, i.e. Cluster.getCenter() will return null.

Parameters:

points - the points to cluster

Returns:

the list of clusters

Throws:

NullArgumentException - if the data points are null