smile.clustering

Class DENCLUE

java.lang.Object

smile.clustering.PartitionClustering

smile.clustering.DENCLUE

All Implemented Interfaces:

java.io.Serializable

public class DENCLUE
extends PartitionClustering

DENsity CLUstering. The DENCLUE algorithm employs a cluster model based on kernel density estimation. A cluster is defined by a local maximum of the estimated density function. Observations going to the same local maximum are put into the same cluster.

Clearly, DENCLUE doesn't work on data with uniform distribution. In high dimensional space, the data always look like uniformly distributed because of the curse of dimensionality. Therefore, DENCLUDE doesn't work well on high-dimensional data in general.

References

1. A. Hinneburg and D. A. Keim. A general approach to clustering in large databases with noise. Knowledge and Information Systems, 5(4):387-415, 2003.
2. Alexander Hinneburg and Hans-Henning Gabriel. DENCLUE 2.0: Fast Clustering based on Kernel Density Estimation. IDA, 2007.

See Also:

Serialized Form

Field Summary

Fields

Modifier and Type	Field and Description
double[][]	attractors The density attractor of each observation.

Fields inherited from class smile.clustering.PartitionClustering

k, OUTLIER, size, y

Constructor Summary

Constructors

Constructor and Description
DENCLUE(int k, double[][] attractors, double[] radius, double[][] samples, double sigma, int[] y, double tol) Constructor.

Method Summary

Modifier and Type	Method and Description
static DENCLUE	fit(double[][] data, double sigma, int m) Clustering data.
static DENCLUE	fit(double[][] data, double sigma, int m, double tol, int minPts) Clustering data.
int	predict(double[] x) Classifies a new observation.

Methods inherited from class smile.clustering.PartitionClustering

run, seed, toString

Methods inherited from class java.lang.Object

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

Field Detail

attractors

public final double[][] attractors

The density attractor of each observation.

Constructor Detail

DENCLUE

public DENCLUE(int k,
               double[][] attractors,
               double[] radius,
               double[][] samples,
               double sigma,
               int[] y,
               double tol)

Constructor.

Parameters:

k - the number of clusters.

attractors - the density attractor of each observation.

radius - the radius of density attractor.

sigma - the smooth parameter in the Gaussian kernel. The user can choose sigma such that number of density attractors is constant for a long interval of sigma.

y - the cluster labels.

tol - the tolerance of hill-climbing procedure.

Method Detail

fit

public static DENCLUE fit(double[][] data,
                          double sigma,
                          int m)

Clustering data.

Parameters:

data - the input data of which each row is an observation.

sigma - the smooth parameter in the Gaussian kernel. The user can choose sigma such that number of density attractors is constant for a long interval of sigma.

m - the number of selected samples used in the iteration. This number should be much smaller than the number of observations to speed up the algorithm. It should also be large enough to capture the sufficient information of underlying distribution.

fit

public static DENCLUE fit(double[][] data,
                          double sigma,
                          int m,
                          double tol,
                          int minPts)

Clustering data.

Parameters:

data - the input data of which each row is an observation.

sigma - the smooth parameter in the Gaussian kernel. The user can choose sigma such that number of density attractors is constant for a long interval of sigma.

m - the number of selected samples used in the iteration. This number should be much smaller than the number of observations to speed up the algorithm. It should also be large enough to capture the sufficient information of underlying distribution.

tol - the tolerance of hill-climbing procedure.

minPts - the minimum number of neighbors for a core attractor.

predict

public int predict(double[] x)

Classifies a new observation.

Parameters:

x - a new observation.

Returns:

the cluster label. Note that it may be PartitionClustering.OUTLIER.