Package elki.outlier.intrinsic

Class ISOS<O>

java.lang.Object

elki.outlier.intrinsic.ISOS<O>

Type Parameters:

O - Object type.

All Implemented Interfaces:

Algorithm, OutlierAlgorithm

@Title("ISOS: Intrinsic Stochastic Outlier Selection")
@Reference(authors="Erich Schubert, Michael Gertz",
           title="Intrinsic t-Stochastic Neighbor Embedding for Visualization and Outlier Detection: A Remedy Against the Curse of Dimensionality?",
           booktitle="Proc. Int. Conf. Similarity Search and Applications, SISAP 2017",
           url="https://doi.org/10.1007/978-3-319-68474-1_13",
           bibkey="DBLP:conf/sisap/SchubertG17")
public class ISOS<O>
extends java.lang.Object
implements OutlierAlgorithm

Intrinsic Stochastic Outlier Selection.

Reference:

Erich Schubert, Michael Gertz
Intrinsic t-Stochastic Neighbor Embedding for Visualization and Outlier Detection: A Remedy Against the Curse of Dimensionality?
Proc. Int. Conf. Similarity Search and Applications, SISAP 2017

Since:

0.7.5

Author:

Erich Schubert

Nested Class Summary

Nested classes/interfaces inherited from interface elki.Algorithm

Algorithm.Utils

Field Summary

Fields

Modifier and Type	Field	Description
protected Distance<? super O>	distance	Distance function used.
protected DistanceBasedIntrinsicDimensionalityEstimator	estimator	Estimator of intrinsic dimensionality.
protected int	k	Number of neighbors (not including query point).
private static Logging	LOG	Class logger.
protected double	phi	Expected outlier rate.

Constructor Summary

Constructors

Constructor	Description
ISOS(Distance<? super O> distance, int k, DistanceBasedIntrinsicDimensionalityEstimator estimator)	Constructor.

Method Summary

Modifier and Type	Method	Description
protected static void	adjustDistances(DBIDRef ignore, DoubleDBIDListIter ki, double max, double id, ModifiableDoubleDBIDList dists)
protected double	estimateID(DBIDRef ignore, DoubleDBIDListIter it, double[] p)	Estimate the local intrinsic dimensionality.
TypeInformation[]	getInputTypeRestriction()	Get the input type restriction used for negotiating the data query.
static void	nominateNeighbors(DBIDIter ignore, DBIDArrayIter di, double[] p, double norm, WritableDoubleDataStore scores)	Vote for neighbors not being outliers.
OutlierResult	run(Relation<O> relation)	Run the algorithm.
static DoubleMinMax	transformScores(WritableDoubleDataStore scores, DBIDs ids, double logPerp, double phi)	Transform scores

Methods inherited from class java.lang.Object

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

Methods inherited from interface elki.outlier.OutlierAlgorithm

autorun

Field Detail

LOG

private static final Logging LOG

Class logger.

distance

protected Distance<? super O> distance

Distance function used.

k

protected int k

Number of neighbors (not including query point).

estimator

protected DistanceBasedIntrinsicDimensionalityEstimator estimator

Estimator of intrinsic dimensionality.

phi

protected double phi

Expected outlier rate.

Constructor Detail

ISOS

public ISOS(Distance<? super O> distance,
            int k,
            DistanceBasedIntrinsicDimensionalityEstimator estimator)

Constructor.

Parameters:

distance - Distance function

k - Number of neighbors to consider

estimator - Estimator of intrinsic dimensionality.

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 OutlierResult run(Relation<O> relation)

Run the algorithm.

Parameters:

relation - data relation.

Returns:

outlier detection result

adjustDistances

protected static void adjustDistances(DBIDRef ignore,
                                      DoubleDBIDListIter ki,
                                      double max,
                                      double id,
                                      ModifiableDoubleDBIDList dists)

estimateID

protected double estimateID(DBIDRef ignore,
                            DoubleDBIDListIter it,
                            double[] p)

Estimate the local intrinsic dimensionality.

Parameters:

ignore - Object to ignore

it - Iterator

p - Scratch array

Returns:

ID estimate

nominateNeighbors

public static void nominateNeighbors(DBIDIter ignore,
                                     DBIDArrayIter di,
                                     double[] p,
                                     double norm,
                                     WritableDoubleDataStore scores)

Vote for neighbors not being outliers. The key method of SOS.

Parameters:

ignore - Object to ignore

di - Neighbor object IDs.

p - Probabilities

norm - Normalization factor (1/sum)

scores - Output score storage

transformScores

public static DoubleMinMax transformScores(WritableDoubleDataStore scores,
                                           DBIDs ids,
                                           double logPerp,
                                           double phi)

Transform scores

Parameters:

scores - Scores to transform

ids - DBIDs to process

logPerp - log perplexity

phi - Expected outlier rate

Returns:

Minimum and maximum scores