oracle.spatial.hadoop.vector.cluster.kmeans

Class KMeans

java.lang.Object

oracle.spatial.hadoop.vector.cluster.kmeans.KMeans

public class KMeans
extends java.lang.Object

Contains constants, configuration parameter names and utility methods for KMean clusteting

Author:

hsaucedo

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	KMeans.KMeansCounters Counters incremented by the KMeans algorithm
static class	KMeans.KMeansIterationResult Holds the result of a cluster iteration.

Field Summary

Fields

Modifier and Type	Field and Description
static java.lang.String	CLUSTERS_FILE_PREFIX The file name prefix of the intermediate clusters file generated between iterations
static java.lang.String	CONF_CLUSTERS_FILE Job configuration: the name of the file that will contain the clusters information between iterations
static java.lang.String	CONF_CRIT_FUN_CLASS Job configuration: the CriterionFunction implementation class that should be used
static java.lang.String	CONF_CRIT_FUN_TERMINATION_VALUE Job configuration: the minimum double value used to determine if a criterion function is converging
static java.lang.String	CONF_K Job configuration: the number of cluster to generate as an integer value
static java.lang.String	CONF_MAX_MEMBER_DISTANCE Maximum distance between a cluster center and a cluster member.
static java.lang.String	CONF_SHAPE_GEN_CLASS Job configuration: the ClusterShapeGenerator implementation class used
static java.lang.String	JSON_CLUSTER_FILE The name of the resulting JSON file
static java.lang.String	MR_OUT_DIR The name of the directory where the intermediate results are stored

Constructor Summary

Constructors

Constructor and Description
KMeans()

Method Summary

Modifier and Type	Method and Description
static Path	buildClustersFilePath(Path outPath, int iteration) Gets the path where the results of the given iteration can be stored
static ClusterInfo[]	readClusters(FileStatus[] clusterSeqFiles, int k, Configuration conf) Reads the information from a list of cluster files
static void	writeClusters(Path clustersFile, ClusterInfo[] clusters, Configuration conf) Writes the given ClusterInfo array to a sequence file
static void	writeJsonClusters(Path clustersFile, ClusterInfo[] clusters, Configuration conf) Writes the given ClusterInfo array to a JSON file

Methods inherited from class java.lang.Object

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

Field Detail

CLUSTERS_FILE_PREFIX

public static final java.lang.String CLUSTERS_FILE_PREFIX

The file name prefix of the intermediate clusters file generated between iterations

See Also:

Constant Field Values

CONF_CLUSTERS_FILE

public static final java.lang.String CONF_CLUSTERS_FILE

Job configuration: the name of the file that will contain the clusters information between iterations

See Also:

Constant Field Values

CONF_CRIT_FUN_CLASS

public static final java.lang.String CONF_CRIT_FUN_CLASS

Job configuration: the CriterionFunction implementation class that should be used

See Also:

Constant Field Values

CONF_CRIT_FUN_TERMINATION_VALUE

public static final java.lang.String CONF_CRIT_FUN_TERMINATION_VALUE

Job configuration: the minimum double value used to determine if a criterion function is converging

See Also:

Constant Field Values

CONF_K

public static final java.lang.String CONF_K

Job configuration: the number of cluster to generate as an integer value

See Also:

Constant Field Values

CONF_MAX_MEMBER_DISTANCE

public static final java.lang.String CONF_MAX_MEMBER_DISTANCE

Maximum distance between a cluster center and a cluster member.

See Also:

Constant Field Values

CONF_SHAPE_GEN_CLASS

public static final java.lang.String CONF_SHAPE_GEN_CLASS

Job configuration: the ClusterShapeGenerator implementation class used

See Also:

Constant Field Values

JSON_CLUSTER_FILE

public static final java.lang.String JSON_CLUSTER_FILE

The name of the resulting JSON file

See Also:

Constant Field Values

MR_OUT_DIR

public static final java.lang.String MR_OUT_DIR

The name of the directory where the intermediate results are stored

See Also:

Constant Field Values

Constructor Detail

KMeans

public KMeans()

Method Detail

buildClustersFilePath

public static Path buildClustersFilePath(Path outPath,
                                         int iteration)

Gets the path where the results of the given iteration can be stored

Parameters:

outPath - the current K-means clustering out path

iteration - the current iteration number

Returns:

readClusters

public static ClusterInfo[] readClusters(FileStatus[] clusterSeqFiles,
                                         int k,
                                         Configuration conf)
                                  throws java.io.IOException

Reads the information from a list of cluster files

Parameters:

clusterSeqFiles - an array of cluster files FileStatus

k - the number of clusters

conf - the job configuration

Returns:

an array of ClusterInfo instances

Throws:

java.io.IOException

writeClusters

public static void writeClusters(Path clustersFile,
                                 ClusterInfo[] clusters,
                                 Configuration conf)
                          throws java.io.IOException

Writes the given ClusterInfo array to a sequence file

Parameters:

clustersFile - the sequence file path where the clusters will be written to

clusters - an array of ClusterInfo instances to be written

conf - the job configuration

Throws:

java.io.IOException

writeJsonClusters

public static void writeJsonClusters(Path clustersFile,
                                     ClusterInfo[] clusters,
                                     Configuration conf)
                              throws java.io.IOException

Writes the given ClusterInfo array to a JSON file

Parameters:

clustersFile - the JSON file path where the clusters will be written to

clusters - an array of ClusterInfo instances to be written

conf - the job configuration

Throws:

java.io.IOException