Package elki.algorithm

Class DependencyDerivator<V extends NumberVector>

java.lang.Object

elki.algorithm.DependencyDerivator<V>

Type Parameters:

V - the type of FeatureVector handled by this Algorithm

All Implemented Interfaces:

Algorithm

@Title("Dependency Derivator: Deriving numerical inter-dependencies on data")
@Description("Derives an equality-system describing dependencies between attributes in a correlation-cluster")
@Reference(authors="Elke Achtert, Christian B\u00f6hm, Hans-Peter Kriegel, Peer Kr\u00f6ger, Arthur Zimek",
           title="Deriving Quantitative Dependencies for Correlation Clusters",
           booktitle="Proc. 12th Int. Conf. on Knowledge Discovery and Data Mining (KDD \'06)",
           url="https://doi.org/10.1145/1150402.1150408",
           bibkey="DBLP:conf/kdd/AchtertBKKZ06")
@Priority(-5)
public class DependencyDerivator<V extends NumberVector>
extends java.lang.Object
implements Algorithm

Dependency derivator computes quantitatively linear dependencies among attributes of a given dataset based on a linear correlation PCA.

Reference:

Elke Achtert, Christian Böhm, Hans-Peter Kriegel, Peer Kröger, Arthur Zimek
Deriving Quantitative Dependencies for Correlation Clusters
Proc. 12th Int. Conf. on Knowledge Discovery and Data Mining (KDD '06)

Since:

0.1

Author:

Arthur Zimek

Nested Class Summary

Nested Classes

Modifier and Type	Class	Description
static class	DependencyDerivator.Par<V extends NumberVector>	Parameterization class.

Nested classes/interfaces inherited from interface elki.Algorithm

Algorithm.Utils

Field Summary

Fields

Modifier and Type	Field	Description
private NumberVectorDistance<? super V>	distance	Distance function used.
private EigenPairFilter	filter	Filter to select eigenvectors.
private static Logging	LOG	The logger for this class.
private java.text.NumberFormat	nf	Number format for output of solution.
private PCARunner	pca	Holds the object performing the pca.
private boolean	randomsample	Flag for random sampling vs. kNN
private int	sampleSize	The number of samples to draw.

Constructor Summary

Constructors

Constructor	Description
DependencyDerivator(NumberVectorDistance<? super V> distance, java.text.NumberFormat nf, PCARunner pca, EigenPairFilter filter, int sampleSize, boolean randomsample)	Constructor.

Method Summary

Modifier and Type	Method	Description
CorrelationAnalysisSolution	generateModel(Relation<V> db, DBIDs ids)	Runs the pca on the given set of IDs.
CorrelationAnalysisSolution	generateModel(Relation<V> relation, DBIDs ids, double[] centroid)	Runs the pca on the given set of IDs and for the given centroid.
TypeInformation[]	getInputTypeRestriction()	Get the input type restriction used for negotiating the data query.
CorrelationAnalysisSolution	run(Relation<V> relation)	Computes quantitatively linear dependencies among the attributes of the given database based on a linear correlation PCA.

Methods inherited from class java.lang.Object

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

Methods inherited from interface elki.Algorithm

autorun

Field Detail

LOG

private static final Logging LOG

The logger for this class.

distance

private NumberVectorDistance<? super V extends NumberVector> distance

Distance function used.

sampleSize

private final int sampleSize

The number of samples to draw.

pca

private final PCARunner pca

Holds the object performing the pca.

filter

private final EigenPairFilter filter

Filter to select eigenvectors.

nf

private final java.text.NumberFormat nf

Number format for output of solution.

randomsample

private final boolean randomsample

Flag for random sampling vs. kNN

Constructor Detail

DependencyDerivator

public DependencyDerivator(NumberVectorDistance<? super V> distance,
                           java.text.NumberFormat nf,
                           PCARunner pca,
                           EigenPairFilter filter,
                           int sampleSize,
                           boolean randomsample)

Constructor.

Parameters:

distance - distance function

nf - Number format

pca - PCA runner

filter - Eigenvector filter

sampleSize - sample size

randomsample - flag for random sampling

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 CorrelationAnalysisSolution run(Relation<V> relation)

Computes quantitatively linear dependencies among the attributes of the given database based on a linear correlation PCA.

Parameters:

relation - the relation to process

Returns:

the CorrelationAnalysisSolution computed by this DependencyDerivator

generateModel

public CorrelationAnalysisSolution generateModel(Relation<V> db,
                                                 DBIDs ids)

Runs the pca on the given set of IDs. The centroid is computed from the given ids.

Parameters:

db - the database

ids - the set of ids

Returns:

a matrix of equations describing the dependencies

generateModel

public CorrelationAnalysisSolution generateModel(Relation<V> relation,
                                                 DBIDs ids,
                                                 double[] centroid)

Runs the pca on the given set of IDs and for the given centroid.

Parameters:

relation - the database

ids - the set of ids

centroid - the centroid

Returns:

a matrix of equations describing the dependencies