de.lmu.ifi.dbs.elki.algorithm.outlier.distance

Class HilOut.HilFeature

java.lang.Object

de.lmu.ifi.dbs.elki.algorithm.outlier.distance.HilOut.HilFeature

All Implemented Interfaces:

java.lang.Comparable<HilOut.HilFeature>

Enclosing class:

HilOut<O extends NumberVector>

static final class HilOut.HilFeature
extends java.lang.Object
implements java.lang.Comparable<HilOut.HilFeature>

Hilbert representation of a single object. Details of this representation are discussed in the main HilOut publication, see "point features".

Author:

Jonathan von Brünken

Field Summary

Fields

Modifier and Type	Field and Description
long[]	hilbert Hilbert representation TODO: use byte[] to save some memory, but slower?
DBID	id Object ID
double	lbound Lower bound of object
int	level Object level
ObjectHeap<DoubleDBIDPair>	nn Heap with the nearest known neighbors
HashSetModifiableDBIDs	nn_keys Set representation of the nearest neighbors for faster lookups
double	sum_nn Current weight (sum of nn distances)
double	ubound Upper bound for object

Constructor Summary

Constructors

Constructor and Description
HilFeature(DBID id, ObjectHeap<DoubleDBIDPair> nn) Constructor.

Method Summary

Modifier and Type	Method and Description
int	compareTo(HilOut.HilFeature o)
protected void	insert(DBID id, double dt, int k) insert function inserts a nearest neighbor into a features nn list and its distance

Methods inherited from class java.lang.Object

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

Field Detail

id

public DBID id

Object ID

hilbert

public long[] hilbert

Hilbert representation TODO: use byte[] to save some memory, but slower?

level

public int level

Object level

ubound

public double ubound

Upper bound for object

lbound

public double lbound

Lower bound of object

nn

public ObjectHeap<DoubleDBIDPair> nn

Heap with the nearest known neighbors

nn_keys

public HashSetModifiableDBIDs nn_keys

Set representation of the nearest neighbors for faster lookups

sum_nn

public double sum_nn

Current weight (sum of nn distances)

Constructor Detail

HilFeature

public HilFeature(DBID id,
                  ObjectHeap<DoubleDBIDPair> nn)

Constructor.

Parameters:

id - Object ID

nn - Heap for neighbors

Method Detail

compareTo

public int compareTo(HilOut.HilFeature o)

Specified by:

compareTo in interface java.lang.Comparable<HilOut.HilFeature>

insert

protected void insert(DBID id,
                      double dt,
                      int k)

insert function inserts a nearest neighbor into a features nn list and its distance

Parameters:

id - DBID of the nearest neighbor

dt - distance or the neighbor to the features position

k - K