de.lmu.ifi.dbs.elki.index.vafile

Class VALPNormDistance

java.lang.Object

de.lmu.ifi.dbs.elki.index.vafile.VALPNormDistance

public class VALPNormDistance
extends java.lang.Object

Lp-Norm distance function for partially computed objects.

Since:

0.5.0

Author:

Erich Schubert

Field Summary

Fields

Modifier and Type	Field and Description
private double[][]	lookup Lookup table for grid cells.
private double	onebyp Value of 1/p for lP norm.
private VectorApproximation	queryApprox Approximation of the query vector.

Constructor Summary

Constructors

Constructor and Description
VALPNormDistance(double p, double[][] splitPositions, NumberVector query, VectorApproximation queryApprox) Constructor.

Method Summary

Modifier and Type	Method and Description
double	getMaxDist(VectorApproximation vec) Get the maximum distance.
double	getMinDist(VectorApproximation vec) Get the minimum distance to approximated vector vec.
double	getPartialMaxDist(int dimension, int vp) Get the maximum distance contribution of a single dimension.
double	getPartialMaxMaxDist(int dimension) Get the maximum distance.
double	getPartialMinDist(int dimension, int vp) Get the minimum distance contribution of a single dimension.
private void	initializeLookupTable(double[][] splitPositions, NumberVector query, double p) Initialize the lookup table.

Methods inherited from class java.lang.Object

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

Field Detail

onebyp

private final double onebyp

Value of 1/p for lP norm.

lookup

private double[][] lookup

Lookup table for grid cells.

queryApprox

private VectorApproximation queryApprox

Approximation of the query vector.

Constructor Detail

VALPNormDistance

public VALPNormDistance(double p,
                        double[][] splitPositions,
                        NumberVector query,
                        VectorApproximation queryApprox)

Constructor.

Parameters:

p - Value of p

splitPositions - Split positions

query - Query vector

queryApprox - Query approximation

Method Detail

getPartialMinDist

public double getPartialMinDist(int dimension,
                                int vp)

Get the minimum distance contribution of a single dimension.

Parameters:

dimension - Dimension

vp - Vector position

Returns:

Increment

getMinDist

public double getMinDist(VectorApproximation vec)

Get the minimum distance to approximated vector vec.

Parameters:

vec - Vector approximation

Returns:

Minimum distance

getPartialMaxDist

public double getPartialMaxDist(int dimension,
                                int vp)

Get the maximum distance contribution of a single dimension.

Parameters:

dimension - Dimension

vp - Vector position

Returns:

Increment

getMaxDist

public double getMaxDist(VectorApproximation vec)

Get the maximum distance.

Parameters:

vec - Approximation vector

Returns:

Maximum distance of the vector

getPartialMaxMaxDist

public double getPartialMaxMaxDist(int dimension)

Get the maximum distance.

Parameters:

dimension - Dimension

Returns:

Maximum distance in the given dimension

initializeLookupTable

private void initializeLookupTable(double[][] splitPositions,
                                   NumberVector query,
                                   double p)

Initialize the lookup table.

Parameters:

splitPositions - Split positions

query - Query vector

p - p