de.lmu.ifi.dbs.elki.algorithm.clustering.hierarchical.extraction

Class HDBSCANHierarchyExtraction.TempCluster

java.lang.Object

de.lmu.ifi.dbs.elki.algorithm.clustering.hierarchical.extraction.HDBSCANHierarchyExtraction.TempCluster

Enclosing class:

HDBSCANHierarchyExtraction

protected static class HDBSCANHierarchyExtraction.TempCluster
extends Object

Temporary cluster.

Author:

Erich Schubert

Field Summary

Fields

Modifier and Type	Field and Description
protected double	aggregate Mass aggregate.
protected Collection<HDBSCANHierarchyExtraction.TempCluster>	children (Finished) child clusters
protected int	childrenTotal Number of objects in children.
protected double	dist Current height.
protected ModifiableDBIDs	members New ids, not yet in child clusters.

Constructor Summary

Constructors

Constructor and Description
HDBSCANHierarchyExtraction.TempCluster(double dist) Constructor.
HDBSCANHierarchyExtraction.TempCluster(double dist, DBIDRef a) Constructor.
HDBSCANHierarchyExtraction.TempCluster(double dist, DBIDRef a, DBIDRef b) Constructor.
HDBSCANHierarchyExtraction.TempCluster(double dist, HDBSCANHierarchyExtraction.TempCluster a, HDBSCANHierarchyExtraction.TempCluster b) Cluster containing two existing clusters.

Method Summary

Methods

Modifier and Type	Method and Description
private void	collectChildren(Clustering<DendrogramModel> clustering, HDBSCANHierarchyExtraction.TempCluster cur, Cluster<DendrogramModel> clus, boolean flatten, boolean hierarchical) Recursive flattening of clusters.
double	excessOfMass() Excess of mass measure.
private void	finalizeCluster(Clustering<DendrogramModel> clustering, Cluster<DendrogramModel> parent, boolean flatten, boolean hierarchical) Make the cluster for the given object
HDBSCANHierarchyExtraction.TempCluster	grow(double dist, DBIDRef id) Add new objects to the cluster.
HDBSCANHierarchyExtraction.TempCluster	grow(double dist, HDBSCANHierarchyExtraction.TempCluster other, DBIDRef id) Join the contents of another cluster.
boolean	isSpurious(int minClSize) Test if a cluster is spurious.
HDBSCANHierarchyExtraction.TempCluster	resetAggregate() Reset the aggregate (for spurious clusters).
int	totalElements() Total number of elements in this subtree.
double	totalStability() Excess of mass measure.

Methods inherited from class java.lang.Object

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

Field Detail

members

protected ModifiableDBIDs members

New ids, not yet in child clusters.

dist

protected double dist

Current height.

aggregate

protected double aggregate

Mass aggregate.

childrenTotal

protected int childrenTotal

Number of objects in children.

children

protected Collection<HDBSCANHierarchyExtraction.TempCluster> children

(Finished) child clusters

Constructor Detail

HDBSCANHierarchyExtraction.TempCluster

public HDBSCANHierarchyExtraction.TempCluster(double dist)

Constructor.

Parameters:

dist - Distance

HDBSCANHierarchyExtraction.TempCluster

public HDBSCANHierarchyExtraction.TempCluster(double dist,
                                      DBIDRef a)

Constructor.

Parameters:

dist - Distance

a - Object reference

HDBSCANHierarchyExtraction.TempCluster

public HDBSCANHierarchyExtraction.TempCluster(double dist,
                                      DBIDRef a,
                                      DBIDRef b)

Constructor.

Parameters:

dist - Distance

HDBSCANHierarchyExtraction.TempCluster

public HDBSCANHierarchyExtraction.TempCluster(double dist,
                                      HDBSCANHierarchyExtraction.TempCluster a,
                                      HDBSCANHierarchyExtraction.TempCluster b)

Cluster containing two existing clusters.

Parameters:

dist - Distance

a - First cluster

b - Second cluster

Method Detail

grow

public HDBSCANHierarchyExtraction.TempCluster grow(double dist,
                                          HDBSCANHierarchyExtraction.TempCluster other,
                                          DBIDRef id)

Join the contents of another cluster.

Parameters:

dist - Join distance

other - Other cluster (may be null)

id - Cluster lead, for 1-element clusters.

Returns:

this

grow

public HDBSCANHierarchyExtraction.TempCluster grow(double dist,
                                          DBIDRef id)

Add new objects to the cluster.

Parameters:

dist - Distance

id - Object to add

Returns:

this

resetAggregate

public HDBSCANHierarchyExtraction.TempCluster resetAggregate()

Reset the aggregate (for spurious clusters).

Returns:

this

totalElements

public int totalElements()

Total number of elements in this subtree.

Returns:

Total

excessOfMass

public double excessOfMass()

Excess of mass measure.

Returns:

Excess of mass

totalStability

public double totalStability()

Excess of mass measure.

Returns:

Excess of mass

isSpurious

public boolean isSpurious(int minClSize)

Test if a cluster is spurious.

Parameters:

minClSize - Minimum cluster size

Returns:

true if spurious

finalizeCluster

private void finalizeCluster(Clustering<DendrogramModel> clustering,
                   Cluster<DendrogramModel> parent,
                   boolean flatten,
                   boolean hierarchical)

Make the cluster for the given object

Parameters:

clustering - Parent clustering

parent - Parent cluster (for hierarchical output)

flatten - Flag to flatten all clusters below.

hierarchical - Hierarchical outpu

collectChildren

private void collectChildren(Clustering<DendrogramModel> clustering,
                   HDBSCANHierarchyExtraction.TempCluster cur,
                   Cluster<DendrogramModel> clus,
                   boolean flatten,
                   boolean hierarchical)

Recursive flattening of clusters.

Parameters:

clustering - Output clustering

cur - Current temporary cluster

clus - Output cluster

flatten - Flag to indicate everything below should be flattened.

hierarchical - Hierarchical output