Package elki.application.benchmark

Class KNNBenchmark<O>

java.lang.Object

elki.application.AbstractApplication

elki.application.AbstractDistanceBasedApplication<O>

elki.application.benchmark.KNNBenchmark<O>

Type Parameters:

O - Object type

public class KNNBenchmark<O>
extends AbstractDistanceBasedApplication<O>

Benchmarking experiment that computes the k nearest neighbors for each query point. The query points can either come from a separate data source, or from the original database.

Since:

0.5.5

Author:

Erich Schubert

Field Summary

Fields

Modifier and Type	Field	Description
protected int	k	Number of neighbors to retrieve.
private static Logging	LOG	The logger for this class.
protected DatabaseConnection	queries	The alternate query point source.
protected RandomFactory	random	Random generator factory
protected double	sampling	Sampling size.

Fields inherited from class elki.application.AbstractDistanceBasedApplication

distance, inputstep

Fields inherited from class elki.application.AbstractApplication

REFERENCE, VERSION

Constructor Summary

Constructors

Constructor	Description
KNNBenchmark(InputStep inputstep, Distance<? super O> distance, int k, DatabaseConnection queries, double sampling, RandomFactory random)	Constructor.

Method Summary

Modifier and Type	Method	Description
private void	logIndexStatistics(Database database)	Log index statistics before and after querying.
static void	main(java.lang.String[] args)	Runs the benchmark
void	run()	Runs the application.
private int	run(KNNSearcher<DBIDRef> knnQuery, Relation<O> relation, Duration dur, MeanVariance mv, MeanVariance mvdist)	Run with the database as query source
private int	run(KNNSearcher<O> knnQuery, Duration dur, MeanVariance mv, MeanVariance mvdist)	Run using a second database as query source

Methods inherited from class elki.application.AbstractApplication

printErrorMessage, runCLIApplication, usage

Methods inherited from class java.lang.Object

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

Field Detail

LOG

private static final Logging LOG

The logger for this class.

k

protected int k

Number of neighbors to retrieve.

queries

protected DatabaseConnection queries

The alternate query point source. Optional.

sampling

protected double sampling

Sampling size.

random

protected RandomFactory random

Random generator factory

Constructor Detail

KNNBenchmark

public KNNBenchmark(InputStep inputstep,
                    Distance<? super O> distance,
                    int k,
                    DatabaseConnection queries,
                    double sampling,
                    RandomFactory random)

Constructor.

Parameters:

inputstep - Input step

distance - Distance function to use

k - K parameter

queries - Query data set (may be null!)

sampling - Sampling rate

random - Random factory

Method Detail

run

public void run()

Description copied from class: AbstractApplication

Runs the application.

Specified by:

run in class AbstractApplication

logIndexStatistics

private void logIndexStatistics(Database database)

Log index statistics before and after querying.

Parameters:

database - Database

run

private int run(KNNSearcher<DBIDRef> knnQuery,
                Relation<O> relation,
                Duration dur,
                MeanVariance mv,
                MeanVariance mvdist)

Run with the database as query source

Parameters:

knnQuery - Query object

relation - Input data

dur - Duration

mv - statistics collector

mvdist - statistics collector

Returns:

hash code of the results

run

private int run(KNNSearcher<O> knnQuery,
                Duration dur,
                MeanVariance mv,
                MeanVariance mvdist)

Run using a second database as query source

Parameters:

knnQuery - Query object

dur - Duration

mv - statistics collector

mvdist - statistics collector

Returns:

hash code of the results

main

public static void main(java.lang.String[] args)

Runs the benchmark

Parameters:

args - parameter list according to description