org.openimaj.hadoop.tools.fastkmeans

Class AKMeans

java.lang.Object

org.openimaj.hadoop.tools.fastkmeans.AKMeans

public class AKMeans
extends Object

Approximate KMeans mapreduce implementation

Author:

Sina Samangooei (ss@ecs.soton.ac.uk)

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	AKMeans.Combine for efficiency, combine centroids early, emitting sums and k for centroids combined
static class	AKMeans.Map the map for approximate kmeans.
static class	AKMeans.Reduce The AKmeans reducer.

Field Summary

Fields

Modifier and Type	Field and Description
static String	CENTROIDS_EXACT Config option where for exact mode or not
static String	CENTROIDS_K Config option where for number of centroids K
static String	CENTROIDS_PATH Config option where for centroids path

Constructor Summary

Constructors

Constructor and Description
AKMeans()

Method Summary

Modifier and Type	Method and Description
static ByteCentroidsResult	completeCentroids(String centroids, String selected, HadoopFastKMeansOptions options) Given the location of a binary dump of centroids on the HDFS, load the binary dump and construct a proper ByteKMeans instance
static ByteCentroidsResult	sequenceFileToCluster(String initialCentroids, int k) load some initially selected centroids from FeatureSelect as a ByteKMeans instance

Methods inherited from class java.lang.Object

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

Field Detail

CENTROIDS_PATH

public static final String CENTROIDS_PATH

Config option where for centroids path

See Also:

Constant Field Values

CENTROIDS_K

public static final String CENTROIDS_K

Config option where for number of centroids K

See Also:

Constant Field Values

CENTROIDS_EXACT

public static final String CENTROIDS_EXACT

Config option where for exact mode or not

See Also:

Constant Field Values

Constructor Detail

AKMeans

public AKMeans()

Method Detail

completeCentroids

public static ByteCentroidsResult completeCentroids(String centroids,
                                                    String selected,
                                                    HadoopFastKMeansOptions options)
                                             throws Exception

Given the location of a binary dump of centroids on the HDFS, load the binary dump and construct a proper ByteKMeans instance

Parameters:

centroids -

selected -

options -

Returns:

ByteKMeans for the centoirds on the HDFS

Throws:

Exception

sequenceFileToCluster

public static ByteCentroidsResult sequenceFileToCluster(String initialCentroids,
                                                        int k)
                                                 throws IOException

load some initially selected centroids from FeatureSelect as a ByteKMeans instance

Parameters:

initialCentroids -

k -

Returns:

a ByteKMeans

Throws:

IOException