tutorial.clustering

Class SameSizeKMeansAlgorithm.Meta

java.lang.Object

tutorial.clustering.SameSizeKMeansAlgorithm.Meta

Enclosing class:

SameSizeKMeansAlgorithm<V extends NumberVector<?>>

private class SameSizeKMeansAlgorithm.Meta
extends Object

Object metadata.

Field Summary

Fields

Modifier and Type	Field and Description
(package private) double[]	dists Distance to the cluster centers.
(package private) int	primary Indexes: primary assignment (current or best), secondary assignment (second best or worst).
(package private) int	secondary Indexes: primary assignment (current or best), secondary assignment (second best or worst).

Constructor Summary

Constructors

Modifier	Constructor and Description
protected	SameSizeKMeansAlgorithm.Meta(int k) Constructor.

Method Summary

Methods

Modifier and Type	Method and Description
protected double	gain(int i) Gain from switching to cluster i.
protected double	priority() Priority / badness: difference between best and worst.

Methods inherited from class java.lang.Object

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

Field Detail

dists

double[] dists

Distance to the cluster centers.

primary

int primary

Indexes: primary assignment (current or best), secondary assignment (second best or worst). The actual meanin changes from initialization to iteration phase!

secondary

int secondary

Indexes: primary assignment (current or best), secondary assignment (second best or worst). The actual meanin changes from initialization to iteration phase!

Constructor Detail

SameSizeKMeansAlgorithm.Meta

protected SameSizeKMeansAlgorithm.Meta(int k)

Constructor.

Parameters:

k -

Method Detail

priority

protected double priority()

Priority / badness: difference between best and worst. (Assuming that "secondary" is the worst).

Returns:

Priority

gain

protected double gain(int i)

Gain from switching to cluster i.

Parameters:

i - Target cluster

Returns:

Gain