Package elki.clustering.hierarchical.extraction

Class HDBSCANHierarchyExtraction.Instance

java.lang.Object

elki.clustering.hierarchical.extraction.HDBSCANHierarchyExtraction.Instance

Enclosing class:

HDBSCANHierarchyExtraction

protected class HDBSCANHierarchyExtraction.Instance
extends java.lang.Object

Instance for a single data set.

Author:

Erich Schubert

Field Summary

Fields

Modifier and Type	Field	Description
protected DoubleDataStore	coredist	Core distances, if available.
protected ClusterMergeHistory	merges	The hierarchical result to process.

Constructor Summary

Constructors

Constructor	Description
Instance(ClusterMergeHistory merges)	Constructor.

Method Summary

Modifier and Type	Method	Description
private double	collectChildren(HDBSCANHierarchyExtraction.TempCluster temp, Clustering<DendrogramModel> clustering, WritableDoubleDataStore glosh, HDBSCANHierarchyExtraction.TempCluster cur, Cluster<DendrogramModel> clus, boolean flatten)	Recursive flattening of clusters.
private double	finalizeCluster(HDBSCANHierarchyExtraction.TempCluster temp, Clustering<DendrogramModel> clustering, WritableDoubleDataStore glosh, Cluster<DendrogramModel> parent, boolean flatten)	Make the cluster for the given object
private boolean	isSpurious(HDBSCANHierarchyExtraction.TempCluster clus, boolean isCore)	Spurious, also for non-materialized clusters.
Clustering<DendrogramModel>	run()	Extract all clusters from the pi-lambda-representation.

Methods inherited from class java.lang.Object

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

Field Detail

merges

protected ClusterMergeHistory merges

The hierarchical result to process.

coredist

protected DoubleDataStore coredist

Core distances, if available.

Constructor Detail

Instance

public Instance(ClusterMergeHistory merges)

Constructor.

Parameters:

merges - Hierarchical result

Method Detail

run

public Clustering<DendrogramModel> run()

Extract all clusters from the pi-lambda-representation.

Returns:

Hierarchical clustering

isSpurious

private boolean isSpurious(HDBSCANHierarchyExtraction.TempCluster clus,
                           boolean isCore)

Spurious, also for non-materialized clusters.

Parameters:

clus - Cluster, may be null for 1-element clusters.

isCore - Core property

Returns:

true if spurious.

finalizeCluster

private double finalizeCluster(HDBSCANHierarchyExtraction.TempCluster temp,
                               Clustering<DendrogramModel> clustering,
                               WritableDoubleDataStore glosh,
                               Cluster<DendrogramModel> parent,
                               boolean flatten)

Make the cluster for the given object

Parameters:

temp - Current temporary cluster

clustering - Parent clustering

glosh - GLOSH scores output

parent - Parent cluster (for hierarchical output)

flatten - Flag to flatten all clusters below

Returns:

smallest distance when the cluster exists

collectChildren

private double collectChildren(HDBSCANHierarchyExtraction.TempCluster temp,
                               Clustering<DendrogramModel> clustering,
                               WritableDoubleDataStore glosh,
                               HDBSCANHierarchyExtraction.TempCluster cur,
                               Cluster<DendrogramModel> clus,
                               boolean flatten)

Recursive flattening of clusters.

Parameters:

clustering - Output clustering

glosh - GLOSH scores output

cur - Current temporary cluster

clus - Output cluster

flatten - Flag to indicate everything below should be flattened

Returns:

smallest distance when the cluster exists