Package elki.clustering.correlation

Class HiCO

java.lang.Object

elki.clustering.correlation.HiCO

All Implemented Interfaces:

Algorithm, GeneralizedOPTICS, OPTICSTypeAlgorithm

@Title("Mining Hierarchies of Correlation Clusters")
@Description("Algorithm for detecting hierarchies of correlation clusters.")
@Reference(authors="Elke Achtert, Christian B\u00f6hm, Peer Kr\u00f6ger, Arthur Zimek",
           title="Mining Hierarchies of Correlation Clusters",
           booktitle="Proc. Int. Conf. on Scientific and Statistical Database Management (SSDBM\'06)",
           url="https://doi.org/10.1109/SSDBM.2006.35",
           bibkey="DBLP:conf/ssdbm/AchtertBKZ06")
public class HiCO
extends java.lang.Object
implements GeneralizedOPTICS

Implementation of the HiCO algorithm, an algorithm for detecting hierarchies of correlation clusters.

Reference:

Elke Achtert, Christian Böhm, Peer Kröger, Arthur Zimek
Mining Hierarchies of Correlation Clusters.
Proc. Int. Conf. on Scientific and Statistical Database Management (SSDBM'06)

Since:

0.1

Author:

Elke Achtert

Nested Class Summary

Nested Classes

Modifier and Type	Class	Description
private class	HiCO.Instance	Instance of the OPTICS algorithm.
static class	HiCO.Par	Parameterization class.

Nested classes/interfaces inherited from interface elki.Algorithm

Algorithm.Utils

Field Summary

Fields

Modifier and Type	Field	Description
static double	DEFAULT_ALPHA	The default value for HiCO.Par.ALPHA_ID.
static double	DEFAULT_DELTA	The default value for HiCO.Par.DELTA_ID.
private double	deltasq	Delta parameter
private EigenPairFilter	filter	Filter for selecting eigenvectors
private int	k	Number of neighbors to query.
private static Logging	LOG	The logger for this class.
private int	mu	Mu parameter.
private PCARunner	pca	PCA utility object.

Constructor Summary

Constructors

Constructor	Description
HiCO(int k, PCARunner pca, double alpha, int mu, double delta)	Constructor.

Method Summary

Modifier and Type	Method	Description
private void	adjust(double[][] v, double[] vector, int corrDim)	Inserts the specified vector into the given orthonormal matrix v at column corrDim.
int	correlationDistance(PCAFilteredResult pca1, PCAFilteredResult pca2, int dimensionality)	Computes the correlation distance between the two subspaces defined by the specified PCAs.
TypeInformation[]	getInputTypeRestriction()	Get the input type restriction used for negotiating the data query.
int	getMinPts()	Get the minpts value used.
ClusterOrder	run(Relation<? extends NumberVector> relation)	Run the HiCO algorithm.

Methods inherited from class java.lang.Object

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

Methods inherited from interface elki.clustering.optics.OPTICSTypeAlgorithm

autorun

Field Detail

LOG

private static final Logging LOG

The logger for this class.

DEFAULT_DELTA

public static final double DEFAULT_DELTA

The default value for HiCO.Par.DELTA_ID.

See Also:

Constant Field Values

DEFAULT_ALPHA

public static final double DEFAULT_ALPHA

The default value for HiCO.Par.ALPHA_ID.

See Also:

Constant Field Values

deltasq

private double deltasq

Delta parameter

mu

private int mu

Mu parameter.

k

private int k

Number of neighbors to query.

pca

private PCARunner pca

PCA utility object.

filter

private EigenPairFilter filter

Filter for selecting eigenvectors

Constructor Detail

HiCO

public HiCO(int k,
            PCARunner pca,
            double alpha,
            int mu,
            double delta)

Constructor.

Parameters:

k - number of neighbors k

pca - PCA class

alpha - Filter for selecting eigenvectors

mu - Mu parameter

delta - Delta parameter

Method Detail

getInputTypeRestriction

public TypeInformation[] getInputTypeRestriction()

Description copied from interface: Algorithm

Get the input type restriction used for negotiating the data query.

Specified by:

getInputTypeRestriction in interface Algorithm

Returns:

Type restriction

run

public ClusterOrder run(Relation<? extends NumberVector> relation)

Run the HiCO algorithm.

Parameters:

relation - Data relation

Returns:

OPTICS cluster order

correlationDistance

public int correlationDistance(PCAFilteredResult pca1,
                               PCAFilteredResult pca2,
                               int dimensionality)

Computes the correlation distance between the two subspaces defined by the specified PCAs.

Parameters:

pca1 - first PCA

pca2 - second PCA

dimensionality - the dimensionality of the data space

Returns:

the correlation distance between the two subspaces defined by the specified PCAs

adjust

private void adjust(double[][] v,
                    double[] vector,
                    int corrDim)

Inserts the specified vector into the given orthonormal matrix v at column corrDim. After insertion the matrix v is orthonormalized and column corrDim of matrix e_czech is set to the corrDim-th unit vector.

Parameters:

v - the orthonormal matrix of the eigenvectors

vector - the vector to be inserted

corrDim - the column at which the vector should be inserted

getMinPts

public int getMinPts()

Description copied from interface: OPTICSTypeAlgorithm

Get the minpts value used. Needed for OPTICS Xi.

Specified by:

getMinPts in interface OPTICSTypeAlgorithm

Returns:

minpts value