org.openimaj.ml.clustering.kmeans

Class FeatureVectorKMeans.Result<T extends FeatureVector>

java.lang.Object

org.openimaj.ml.clustering.FeatureVectorCentroidsResult<T>

org.openimaj.ml.clustering.kmeans.FeatureVectorKMeans.Result<T>

Type Parameters:

T - Type of object being clustered

All Implemented Interfaces:

Readable, ReadableASCII, ReadableBinary, ReadWriteable, Writeable, WriteableASCII, WriteableBinary, ObjectNearestNeighboursProvider<T>, CentroidsProvider<T>, Clusters, SpatialClusters<T>

Enclosing class:

FeatureVectorKMeans<T extends FeatureVector>

public static class FeatureVectorKMeans.Result<T extends FeatureVector>
extends FeatureVectorCentroidsResult<T>
implements ObjectNearestNeighboursProvider<T>

Result object for FeatureVectorKMeans, extending FeatureVectorCentroidsResult and ObjectNearestNeighboursProvider, as well as giving access to state information from the operation of the K-Means algorithm (i.e. number of iterations, and convergence state).

Author:

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

Field Summary

Fields

Modifier and Type	Field and Description
protected int	changedCentroidCount
protected int	iterations
protected ObjectNearestNeighbours<T>	nn

Fields inherited from class org.openimaj.ml.clustering.FeatureVectorCentroidsResult

centroids

Fields inherited from interface org.openimaj.ml.clustering.Clusters

CLUSTER_HEADER

Constructor Summary

Constructors

Constructor and Description
Result()

Method Summary

Modifier and Type	Method and Description
HardAssigner<T,float[],IntFloatPair>	defaultHardAssigner() Get the default hard assigner for this clusterer.
ObjectNearestNeighbours<T>	getNearestNeighbours()
int	numChangedCentroids() Get the number of changed centroids in the last iteration.
int	numIterations() Get the number of K-Means iterations that produced this result.

Methods inherited from class org.openimaj.ml.clustering.FeatureVectorCentroidsResult

asciiHeader, binaryHeader, equals, getCentroids, numClusters, numDimensions, readASCII, readBinary, toString, writeASCII, writeBinary

Methods inherited from class java.lang.Object

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

Field Detail

nn

protected ObjectNearestNeighbours<T extends FeatureVector> nn

iterations

protected int iterations

changedCentroidCount

protected int changedCentroidCount

Constructor Detail

Result

public Result()

Method Detail

getNearestNeighbours

public ObjectNearestNeighbours<T> getNearestNeighbours()

Specified by:

getNearestNeighbours in interface ObjectNearestNeighboursProvider<T extends FeatureVector>

Returns:

the underlying nearest neighbour object

defaultHardAssigner

public HardAssigner<T,float[],IntFloatPair> defaultHardAssigner()

Description copied from interface: SpatialClusters

Get the default hard assigner for this clusterer. This method is potentially expensive, so callers should only call it once, and hold on to the result (and reuse it).

Specified by:

defaultHardAssigner in interface SpatialClusters<T extends FeatureVector>

Overrides:

defaultHardAssigner in class FeatureVectorCentroidsResult<T extends FeatureVector>

Returns:

a hard assigner.

numIterations

public int numIterations()

Get the number of K-Means iterations that produced this result.

Returns:

the number of iterations

numChangedCentroids

public int numChangedCentroids()

Get the number of changed centroids in the last iteration. This is an indicator of convergence as over time this should reduce to 0.

Returns:

the number of changed centroids