de.lmu.ifi.dbs.elki.distance.distancefunction
Class KernelBasedLocallyWeightedDistanceFunction<O extends RealVector<O,?>,P extends Preprocessor<O>>

java.lang.Object
  de.lmu.ifi.dbs.elki.logging.AbstractLoggable
      de.lmu.ifi.dbs.elki.utilities.optionhandling.AbstractParameterizable
          de.lmu.ifi.dbs.elki.distance.AbstractMeasurementFunction<O,D>
              de.lmu.ifi.dbs.elki.distance.distancefunction.AbstractDistanceFunction<O,DoubleDistance>
                  de.lmu.ifi.dbs.elki.distance.distancefunction.AbstractDoubleDistanceFunction<O>
                      de.lmu.ifi.dbs.elki.distance.distancefunction.AbstractLocallyWeightedDistanceFunction<O,P>
                          de.lmu.ifi.dbs.elki.distance.distancefunction.KernelBasedLocallyWeightedDistanceFunction<O,P>

All Implemented Interfaces:

DistanceFunction<O,DoubleDistance>, MeasurementFunction<O,DoubleDistance>, Loggable, Parameterizable

public class KernelBasedLocallyWeightedDistanceFunction<O extends RealVector<O,?>,P extends Preprocessor<O>>
extends AbstractLocallyWeightedDistanceFunction<O,P>

Provides a kernel based locally weighted distance function. It is defined as follows: result = max{dist_P(P,Q), dist_Q(Q,P)}, where dist_P(P,Q) computes the quadratic form distance on the weak eigenvectors of the kernel matrix of P between two vectors P and Q in feature space. Computation of the distance component of the weak eigenvectors is done indirectly by computing the difference between the complete kernel distance and the distance component of the strong eigenvectors: dist_P(P,Q) = dist_{P_weak}(P,Q) = sqrt(dist_{P_complete}(P,Q)^2 - dist_{P_strong}(P,Q)^2) K_{P_complete}(P,Q) is the kernel derived distance between P and Q. The distance component of the strong eigenvectors K_{P_strong}(P,Q) is computed as follows: First, the vectors P and Q are projected onto the strong eigenvectors of the kernel matrix of P, which results in the two vectors P_p and Q_p. Then, the euklidean distance is used to compute the distance between P_p and Q_p. In case of the linear kernel function, this distance is identical to those computed by the LocallyWeightedDistanceFunction with parameters big = 1.0 and small = 0.0

Author:

Simon Paradies

Field Summary

static String	DEFAULT_KERNEL_FUNCTION_CLASS The default kernel function.
static String	KERNEL_FUNCTION_CLASS_D Description for parameter kernel.
static String	KERNEL_FUNCTION_CLASS_P Parameter for kernel.
private KernelFunction<O,DoubleDistance>	kernelFunction The kernel function that is used.
private KernelMatrix<O>	kernelMatrix The global precomputed kernel matrix

Fields inherited from class de.lmu.ifi.dbs.elki.distance.distancefunction.AbstractLocallyWeightedDistanceFunction

DEFAULT_PREPROCESSOR_CLASS, PREPROCESSOR_CLASS_D

Fields inherited from class de.lmu.ifi.dbs.elki.distance.AbstractMeasurementFunction

INFINITY_PATTERN

Fields inherited from class de.lmu.ifi.dbs.elki.utilities.optionhandling.AbstractParameterizable

optionHandler

Fields inherited from class de.lmu.ifi.dbs.elki.logging.AbstractLoggable

debug

Constructor Summary

KernelBasedLocallyWeightedDistanceFunction()
Provides a kernel based locally weighted distance function.

Method Summary

private double	computeDistance(O o1, O o2) Computes the distance between two given real vectors according to this distance function.
DoubleDistance	distance(O o1, O o2) Computes the distance between two given DatabaseObjects according to this distance function.
(package private) AssociationID	getAssociationID() Returns the assocoiation ID for the association to be set by the preprocessor.
List<AttributeSettings>	getAttributeSettings() Returns the settings of all options assigned to the option handler.
void	setDatabase(Database<O> database, boolean verbose, boolean time) Set the database that holds the associations for the DatabaseObject for which the measurements should be computed.
String[]	setParameters(String[] args) Sets the attributes of the class accordingly to the given parameters.

Methods inherited from class de.lmu.ifi.dbs.elki.distance.distancefunction.AbstractLocallyWeightedDistanceFunction

description

Methods inherited from class de.lmu.ifi.dbs.elki.distance.distancefunction.AbstractDoubleDistanceFunction

infiniteDistance, nullDistance, undefinedDistance, valueOf

Methods inherited from class de.lmu.ifi.dbs.elki.distance.distancefunction.AbstractDistanceFunction

distance, distance, isInfiniteDistance, isNullDistance, isUndefinedDistance

Methods inherited from class de.lmu.ifi.dbs.elki.distance.AbstractMeasurementFunction

getDatabase, matches, requiredInputPattern, setRequiredInputPattern

Methods inherited from class de.lmu.ifi.dbs.elki.utilities.optionhandling.AbstractParameterizable

addOption, checkGlobalParameterConstraints, deleteOption, description, description, getParameters, getParameterValue, getPossibleOptions, inlineDescription, isSet, setParameters

Methods inherited from class de.lmu.ifi.dbs.elki.logging.AbstractLoggable

debugFine, debugFiner, debugFinest, exception, message, progress, progress, progress, verbose, verbose, warning

Methods inherited from class java.lang.Object

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

Methods inherited from interface de.lmu.ifi.dbs.elki.distance.MeasurementFunction

requiredInputPattern

Methods inherited from interface de.lmu.ifi.dbs.elki.utilities.optionhandling.Parameterizable

checkGlobalParameterConstraints, getParameters, getPossibleOptions, inlineDescription

Field Detail

DEFAULT_KERNEL_FUNCTION_CLASS

public static final String DEFAULT_KERNEL_FUNCTION_CLASS

The default kernel function.

KERNEL_FUNCTION_CLASS_D

public static final String KERNEL_FUNCTION_CLASS_D

Description for parameter kernel.

KERNEL_FUNCTION_CLASS_P

public static final String KERNEL_FUNCTION_CLASS_P

Parameter for kernel.

See Also:

Constant Field Values

kernelFunction

private KernelFunction<O extends RealVector<O,?>,DoubleDistance> kernelFunction

The kernel function that is used.

kernelMatrix

private KernelMatrix<O extends RealVector<O,?>> kernelMatrix

The global precomputed kernel matrix

Constructor Detail

KernelBasedLocallyWeightedDistanceFunction

public KernelBasedLocallyWeightedDistanceFunction()

Provides a kernel based locally weighted distance function.

Method Detail

distance

public DoubleDistance distance(O o1,
                               O o2)

Computes the distance between two given DatabaseObjects according to this distance function.

Parameters:

o1 - first DatabaseObject

o2 - second DatabaseObject

Returns:

the distance between two given DatabaseObjects according to this distance function

setParameters

public String[] setParameters(String[] args)
                       throws ParameterException

Description copied from interface: Parameterizable

Sets the attributes of the class accordingly to the given parameters. Returns a new String array containing those entries of the given array that are neither expected nor used by this Parameterizable.

Specified by:

setParameters in interface Parameterizable

Overrides:

setParameters in class AbstractLocallyWeightedDistanceFunction<O extends RealVector<O,?>,P extends Preprocessor<O>>

Parameters:

args - parameters to set the attributes accordingly to

Returns:

String[] an array containing the unused parameters

Throws:

ParameterException - in case of wrong parameter-setting

See Also:

Parameterizable.setParameters(String[])

getAttributeSettings

public List<AttributeSettings> getAttributeSettings()

Description copied from class: AbstractParameterizable

Returns the settings of all options assigned to the option handler.

Specified by:

getAttributeSettings in interface Parameterizable

Overrides:

getAttributeSettings in class AbstractLocallyWeightedDistanceFunction<O extends RealVector<O,?>,P extends Preprocessor<O>>

Returns:

the setting of the attributes of the parameterizable

See Also:

Parameterizable.getAttributeSettings()

computeDistance

private double computeDistance(O o1,
                               O o2)

Computes the distance between two given real vectors according to this distance function.

Parameters:

o1 - first FeatureVector

o2 - second FeatureVector

Returns:

the distance between two given real vectors according to this distance function

setDatabase

public void setDatabase(Database<O> database,
                        boolean verbose,
                        boolean time)

Description copied from interface: MeasurementFunction

Set the database that holds the associations for the DatabaseObject for which the measurements should be computed.

Specified by:

setDatabase in interface MeasurementFunction<O extends RealVector<O,?>,DoubleDistance>

Overrides:

setDatabase in class AbstractLocallyWeightedDistanceFunction<O extends RealVector<O,?>,P extends Preprocessor<O>>

Parameters:

database - the database to be set

verbose - flag to allow verbose messages while performing the method

time - flag to request output of performance time

See Also:

MeasurementFunction.setDatabase(de.lmu.ifi.dbs.elki.database.Database, boolean, boolean)

getAssociationID

AssociationID getAssociationID()

Returns the assocoiation ID for the association to be set by the preprocessor.

Specified by:

getAssociationID in class AbstractLocallyWeightedDistanceFunction<O extends RealVector<O,?>,P extends Preprocessor<O>>