org.openimaj.lsh.functions

Class FloatPStableFactory

java.lang.Object

org.openimaj.lsh.functions.RandomisedHashFunctionFactory<float[]>

org.openimaj.lsh.functions.FloatHashFunctionFactory

org.openimaj.lsh.functions.FloatPStableFactory

All Implemented Interfaces:

HashFunctionFactory<float[]>

Direct Known Subclasses:

FloatCauchyFactory, FloatGaussianFactory

@Reference(type=Inproceedings,
           author={"Datar, Mayur","Immorlica, Nicole","Indyk, Piotr","Mirrokni, Vahab S."},
           title="Locality-sensitive hashing scheme based on p-stable distributions",
           year="2004",
           booktitle="Proceedings of the twentieth annual symposium on Computational geometry",
           pages={"253","","262"},
           url="http://doi.acm.org/10.1145/997817.997857",
           publisher="ACM",
           series="SCG \'04")
public abstract class FloatPStableFactory
extends FloatHashFunctionFactory

Base class for hashing schemes based on P-Stable distributions. The hash functions are of the form h(x) = floor((ax + b) / w).

Author:

Jonathon Hare (jsh2@ecs.soton.ac.uk)

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
protected class	FloatPStableFactory.PStableFunction

Field Summary

Fields inherited from class org.openimaj.lsh.functions.RandomisedHashFunctionFactory

ndims, rng

Constructor Summary

Constructors

Constructor and Description
FloatPStableFactory(int ndims, cern.jet.random.engine.MersenneTwister rng, double w) Construct with the given parameters.

Method Summary

Methods inherited from class org.openimaj.lsh.functions.FloatHashFunctionFactory

distanceFunction, fvDistanceFunction

Methods inherited from class java.lang.Object

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

Methods inherited from interface org.openimaj.util.hash.HashFunctionFactory

create

Constructor Detail

FloatPStableFactory

public FloatPStableFactory(int ndims,
                           cern.jet.random.engine.MersenneTwister rng,
                           double w)

Construct with the given parameters.

Parameters:

ndims - number of dimensions of the data

rng - the random number generator

w - the width parameter