Package elki.clustering.optics

Class FastOPTICS<V extends NumberVector>

java.lang.Object

elki.clustering.optics.FastOPTICS<V>

Type Parameters:

V - Input vector type

All Implemented Interfaces:

Algorithm, OPTICSTypeAlgorithm

@Reference(authors="J. Schneider, M. Vlachos",
           title="Fast parameterless density-based clustering via random projections",
           booktitle="Proc. 22nd ACM Int. Conf. on Information & Knowledge Management (CIKM 2013)",
           url="https://doi.org/10.1145/2505515.2505590",
           bibkey="DBLP:conf/cikm/SchneiderV13")
public class FastOPTICS<V extends NumberVector>
extends java.lang.Object
implements OPTICSTypeAlgorithm

FastOPTICS algorithm (Fast approximation of OPTICS)

Note that this is not FOPTICS as in "Fuzzy OPTICS"!

Reference:

J. Schneider, M. Vlachos
Fast parameterless density-based clustering via random projections
Proc. 22nd ACM Int. Conf. on Information and Knowledge Management (CIKM 2013)

This is based on the original code provided by Johannes Schneider, with ELKIfications and optimizations by Erich Schubert.

Since:

0.7.0

Author:

Johannes Schneider, Erich Schubert

Nested Class Summary

Nested classes/interfaces inherited from interface elki.Algorithm

Algorithm.Utils

Field Summary

Fields

Modifier and Type	Field	Description
(package private) RandomProjectedNeighborsAndDensities	index	Index.
(package private) DoubleDataStore	inverseDensities	Inverse Densities correspond to average distances in point set of projections
private static Logging	LOG	Class logger.
(package private) int	minPts	MinPts parameter.
(package private) DataStore<? extends DBIDs>	neighs	neighbors of a point
(package private) ClusterOrder	order	Result: output order of points
(package private) ModifiableDBIDs	processed	processed points
(package private) WritableDoubleDataStore	reachDist	Result: reachability distances
static double	UNDEFINED_DISTANCE	undefined value for (reachability/average) distance

Constructor Summary

Constructors

Constructor	Description
FastOPTICS(int minpts, RandomProjectedNeighborsAndDensities index)	Constructor.

Method Summary

Modifier and Type	Method	Description
protected void	expandClusterOrder(DBID ipt, ClusterOrder order, DistanceQuery<V> dq, FiniteProgress prog)	OPTICS algorithm for processing a point, but with different density estimates
TypeInformation[]	getInputTypeRestriction()	Get the input type restriction used for negotiating the data query.
int	getMinPts()	Get the minpts value used.
ClusterOrder	run(Relation<V> relation)	Run the 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

Class logger.

UNDEFINED_DISTANCE

public static final double UNDEFINED_DISTANCE

undefined value for (reachability/average) distance

See Also:

Constant Field Values

order

ClusterOrder order

Result: output order of points

reachDist

WritableDoubleDataStore reachDist

Result: reachability distances

processed

ModifiableDBIDs processed

processed points

neighs

DataStore<? extends DBIDs> neighs

neighbors of a point

inverseDensities

DoubleDataStore inverseDensities

Inverse Densities correspond to average distances in point set of projections

minPts

int minPts

MinPts parameter.

index

RandomProjectedNeighborsAndDensities index

Index.

Constructor Detail

FastOPTICS

public FastOPTICS(int minpts,
                  RandomProjectedNeighborsAndDensities index)

Constructor.

Parameters:

minpts - Minimum number of neighbors.

index - Index

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<V> relation)

Run the algorithm.

Parameters:

relation - Relation

expandClusterOrder

protected void expandClusterOrder(DBID ipt,
                                  ClusterOrder order,
                                  DistanceQuery<V> dq,
                                  FiniteProgress prog)

OPTICS algorithm for processing a point, but with different density estimates

Parameters:

ipt - Point

order - Cluster order (output)

dq - Distance query

prog - Progress for logging.

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