de.lmu.ifi.dbs.elki.distance.similarityfunction.kernel

Class PolynomialKernelFunction

java.lang.Object

de.lmu.ifi.dbs.elki.distance.similarityfunction.AbstractPrimitiveSimilarityFunction<NumberVector<?>,DoubleDistance>

de.lmu.ifi.dbs.elki.distance.similarityfunction.AbstractVectorDoubleSimilarityFunction

de.lmu.ifi.dbs.elki.distance.similarityfunction.kernel.PolynomialKernelFunction

All Implemented Interfaces:

DistanceFunction<NumberVector<?>,DoubleDistance>, PrimitiveDistanceFunction<NumberVector<?>,DoubleDistance>, PrimitiveDoubleDistanceFunction<NumberVector<?>>, PrimitiveDoubleSimilarityFunction<NumberVector<?>>, PrimitiveSimilarityFunction<NumberVector<?>,DoubleDistance>, SimilarityFunction<NumberVector<?>,DoubleDistance>, Parameterizable

Direct Known Subclasses:

LinearKernelFunction

public class PolynomialKernelFunction
extends AbstractVectorDoubleSimilarityFunction
implements PrimitiveDoubleDistanceFunction<NumberVector<?>>

Provides a polynomial Kernel function that computes a similarity between the two feature vectors V1 and V2 defined by (V1^T*V2)^degree.

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	PolynomialKernelFunction.Parameterizer Parameterization class.

Field Summary

Fields

Modifier and Type	Field and Description
private double	bias Bias of the similarity function.
static int	DEFAULT_DEGREE The default degree.
private int	degree Degree of the polynomial kernel function.

Constructor Summary

Constructors

Constructor and Description
PolynomialKernelFunction(int degree) Constructor.
PolynomialKernelFunction(int degree, double bias) Constructor.

Method Summary

Methods

Modifier and Type	Method and Description
DoubleDistance	distance(NumberVector<?> fv1, NumberVector<?> fv2) Computes the distance between two given DatabaseObjects according to this distance function.
double	doubleDistance(NumberVector<?> fv1, NumberVector<?> fv2) Computes the distance between two given Objects according to this distance function.
double	doubleSimilarity(NumberVector<?> o1, NumberVector<?> o2) Computes the similarity between two given Objects according to this similarity function.
<T extends NumberVector<?>> DistanceSimilarityQuery<T,DoubleDistance>	instantiate(Relation<T> database) Instantiate with a representation to get the actual similarity query.
boolean	isMetric() Is this distance function metric (in particular, does it satisfy the triangle equation?)

Methods inherited from class de.lmu.ifi.dbs.elki.distance.similarityfunction.AbstractVectorDoubleSimilarityFunction

getDistanceFactory, getInputTypeRestriction, similarity

Methods inherited from class de.lmu.ifi.dbs.elki.distance.similarityfunction.AbstractPrimitiveSimilarityFunction

isSymmetric

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.distancefunction.PrimitiveDistanceFunction

getInputTypeRestriction

Methods inherited from interface de.lmu.ifi.dbs.elki.distance.distancefunction.DistanceFunction

getDistanceFactory, isSymmetric

Methods inherited from interface de.lmu.ifi.dbs.elki.distance.similarityfunction.SimilarityFunction

isSymmetric

Field Detail

DEFAULT_DEGREE

public static final int DEFAULT_DEGREE

The default degree.

See Also:

Constant Field Values

degree

private final int degree

Degree of the polynomial kernel function.

bias

private final double bias

Bias of the similarity function.

Constructor Detail

PolynomialKernelFunction

public PolynomialKernelFunction(int degree,
                        double bias)

Constructor.

Parameters:

degree - Kernel degree

bias - Bias offset

PolynomialKernelFunction

public PolynomialKernelFunction(int degree)

Constructor.

Parameters:

degree - Kernel degree

Method Detail

doubleSimilarity

public double doubleSimilarity(NumberVector<?> o1,
                      NumberVector<?> o2)

Description copied from interface: PrimitiveDoubleSimilarityFunction

Computes the similarity between two given Objects according to this similarity function.

Specified by:

doubleSimilarity in interface PrimitiveDoubleSimilarityFunction<NumberVector<?>>

Parameters:

o1 - first Object

o2 - second Object

Returns:

the similarity between two given Objects according to this similarity function

distance

public DoubleDistance distance(NumberVector<?> fv1,
                      NumberVector<?> fv2)

Description copied from interface: PrimitiveDistanceFunction

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

Specified by:

distance in interface PrimitiveDistanceFunction<NumberVector<?>,DoubleDistance>

Parameters:

fv1 - first DatabaseObject

fv2 - second DatabaseObject

Returns:

the distance between two given DatabaseObjects according to this distance function

isMetric

public boolean isMetric()

Description copied from interface: DistanceFunction

Is this distance function metric (in particular, does it satisfy the triangle equation?)

Specified by:

isMetric in interface DistanceFunction<NumberVector<?>,DoubleDistance>

Returns:

true when metric.

doubleDistance

public double doubleDistance(NumberVector<?> fv1,
                    NumberVector<?> fv2)

Description copied from interface: PrimitiveDoubleDistanceFunction

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

Specified by:

doubleDistance in interface PrimitiveDoubleDistanceFunction<NumberVector<?>>

Parameters:

fv1 - first Object

fv2 - second Object

Returns:

the distance between two given Objects according to this distance function

instantiate

public <T extends NumberVector<?>> DistanceSimilarityQuery<T,DoubleDistance> instantiate(Relation<T> database)

Description copied from interface: SimilarityFunction

Instantiate with a representation to get the actual similarity query.

Specified by:

instantiate in interface DistanceFunction<NumberVector<?>,DoubleDistance>

Specified by:

instantiate in interface SimilarityFunction<NumberVector<?>,DoubleDistance>

Overrides:

instantiate in class AbstractPrimitiveSimilarityFunction<NumberVector<?>,DoubleDistance>

Parameters:

database - Representation to use

Returns:

Actual distance query.