Package elki.clustering.subspace.clique

Class CLIQUEUnit

java.lang.Object

elki.clustering.subspace.clique.CLIQUEUnit

public class CLIQUEUnit
extends java.lang.Object

Represents a unit in the CLIQUE algorithm.

Since:

0.1

Author:

Elke Achtert

Field Summary

Fields

Modifier and Type	Field	Description
private boolean	assigned	Flag that indicates if this unit is already assigned to a cluster.
private double[]	bounds	The bounding values (min, max) for each dimension.
private int[]	dims	The dimensions involved in this subspace.
private ModifiableDBIDs	ids	The ids of the feature vectors this unit contains.

Constructor Summary

Constructors

Modifier	Constructor	Description
	CLIQUEUnit(int dim, double min, double max)	Creates a new one-dimensional unit for the given interval.
private	CLIQUEUnit(CLIQUEUnit prefix, int newdim, double min, double max, ModifiableDBIDs ids)	Creates a new k-dimensional unit for the given intervals.

Method Summary

Modifier and Type	Method	Description
boolean	addFeatureVector(DBIDRef id, NumberVector vector)	Adds the id of the specified feature vector to this unit, if this unit contains the feature vector.
private boolean	checkDimensions(CLIQUEUnit other, int e)	Check that the first e dimensions agree.
boolean	contains(NumberVector vector)	Returns true, if the intervals of this unit contain the specified feature vector.
protected boolean	containsLeftNeighbor(CLIQUEUnit unit, int d)	Returns true if this unit is the left neighbor of the given unit.
protected boolean	containsRightNeighbor(CLIQUEUnit unit, int d)	Returns true if this unit is the right neighbor of the given unit.
int	dimensionality()	Get the dimensionality of this unit.
int	getDimension(int i)	Get the ith dimension constrained.
DBIDs	getIds()	Returns the ids of the feature vectors this unit contains.
boolean	isAssigned()	Returns true if this unit is already assigned to a cluster.
protected CLIQUEUnit	join(CLIQUEUnit other, double all, double tau)	Joins this unit with the specified unit.
void	markAsAssigned()	Marks this unit as assigned to a cluster.
int	numberOfFeatureVectors()	Returns the number of feature vectors this unit contains.
double	selectivity(double total)	Returns the selectivity of this unit, which is defined as the fraction of total feature vectors contained in this unit.
java.lang.String	toString()	Returns a string representation of this unit that contains the intervals of this unit.

Methods inherited from class java.lang.Object

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

Field Detail

dims

private int[] dims

The dimensions involved in this subspace.

bounds

private double[] bounds

The bounding values (min, max) for each dimension.

ids

private ModifiableDBIDs ids

The ids of the feature vectors this unit contains.

assigned

private boolean assigned

Flag that indicates if this unit is already assigned to a cluster.

Constructor Detail

CLIQUEUnit

private CLIQUEUnit(CLIQUEUnit prefix,
                   int newdim,
                   double min,
                   double max,
                   ModifiableDBIDs ids)

Creates a new k-dimensional unit for the given intervals.

Parameters:

prefix - Prefix unit that will be extended by one dimension

newdim - Additional dimension

min - Minimum bound

max - Maximum bound

ids - the ids of the feature vectors belonging to this unit

CLIQUEUnit

public CLIQUEUnit(int dim,
                  double min,
                  double max)

Creates a new one-dimensional unit for the given interval.

Parameters:

dim - Dimension

min - Minimum

max - MAximum

Method Detail

dimensionality

public int dimensionality()

Get the dimensionality of this unit.

Returns:

Number of dimensions constrained.

getDimension

public int getDimension(int i)

Get the ith dimension constrained.

Parameters:

i - Index

Returns:

dimension

contains

public boolean contains(NumberVector vector)

Returns true, if the intervals of this unit contain the specified feature vector.

Parameters:

vector - the feature vector to be tested for containment

Returns:

true, if the intervals of this unit contain the specified feature vector, false otherwise

addFeatureVector

public boolean addFeatureVector(DBIDRef id,
                                NumberVector vector)

Adds the id of the specified feature vector to this unit, if this unit contains the feature vector.

Parameters:

id - Vector id

vector - the feature vector to be added

Returns:

true, if this unit contains the specified feature vector, false otherwise

numberOfFeatureVectors

public int numberOfFeatureVectors()

Returns the number of feature vectors this unit contains.

Returns:

the number of feature vectors this unit contains

selectivity

public double selectivity(double total)

Returns the selectivity of this unit, which is defined as the fraction of total feature vectors contained in this unit.

Parameters:

total - the total number of feature vectors

Returns:

the selectivity of this unit

containsLeftNeighbor

protected boolean containsLeftNeighbor(CLIQUEUnit unit,
                                       int d)

Returns true if this unit is the left neighbor of the given unit.

Parameters:

unit - Reference unit

d - Current dimension

Returns:

true if this unit is the left neighbor of the given unit

containsRightNeighbor

protected boolean containsRightNeighbor(CLIQUEUnit unit,
                                        int d)

Returns true if this unit is the right neighbor of the given unit.

Parameters:

unit - Reference unit

d - Current dimension

Returns:

true if this unit is the right neighbor of the given unit

isAssigned

public boolean isAssigned()

Returns true if this unit is already assigned to a cluster.

Returns:

true if this unit is already assigned to a cluster, false otherwise.

markAsAssigned

public void markAsAssigned()

Marks this unit as assigned to a cluster.

getIds

public DBIDs getIds()

Returns the ids of the feature vectors this unit contains.

Returns:

the ids of the feature vectors this unit contains

join

protected CLIQUEUnit join(CLIQUEUnit other,
                          double all,
                          double tau)

Joins this unit with the specified unit.

Parameters:

other - the unit to be joined

all - the overall number of feature vectors

tau - the density threshold for the selectivity of a unit

Returns:

the joined unit if the selectivity of the join result is equal or greater than tau, null otherwise

checkDimensions

private boolean checkDimensions(CLIQUEUnit other,
                                int e)

Check that the first e dimensions agree.

Parameters:

other - Other unit

e - Number of dimensions to check

Returns:

true if the first e dimensions are the same (index and bounds)

toString

public java.lang.String toString()

Returns a string representation of this unit that contains the intervals of this unit.

Overrides:

toString in class java.lang.Object

Returns:

a string representation of this unit