net.sourceforge.javaocr.plugin.cluster
Class NormalDistributionCluster

java.lang.Object
  net.sourceforge.javaocr.plugin.cluster.AbstractBaseCluster
      net.sourceforge.javaocr.plugin.cluster.EuclidianDistanceCluster
          net.sourceforge.javaocr.plugin.cluster.NormalDistributionCluster

All Implemented Interfaces:

Metric, Cluster

Direct Known Subclasses:

SigmaWeightedEuclidianDistanceCluster

public abstract class NormalDistributionCluster
extends EuclidianDistanceCluster

cluster with normally distributed features. this abstract provides computation of expectation and standard deviation without storing sample values

Author:

Konstantin Pribluda

Field Summary

(package private) double[]	quads
(package private) double[]	var

Constructor Summary

protected	NormalDistributionCluster() default constructor for sake of serialisation frameworks
	NormalDistributionCluster(double[] mx, double[] var) convenience constructor to create already trained distribution cluster
	NormalDistributionCluster(int dimensions) constructs

Method Summary

double[]	getQuads()
double[]	getVar() lazily calculate and return variance cluster
void	setQuads(double[] quads)
void	setVar(double[] var)
void	train(double[] samples) perform sample image training - cumulate values and compute moments

Methods inherited from class net.sourceforge.javaocr.plugin.cluster.EuclidianDistanceCluster

computeDimension, distance

Methods inherited from class net.sourceforge.javaocr.plugin.cluster.AbstractBaseCluster

center, getAmountSamples, getDimensions, getMx, getSum, radius, setAmountSamples, setDimensions, setMx, setSum

Methods inherited from class java.lang.Object

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

Field Detail

quads

double[] quads

var

double[] var

Constructor Detail

NormalDistributionCluster

protected NormalDistributionCluster()

default constructor for sake of serialisation frameworks

NormalDistributionCluster

public NormalDistributionCluster(int dimensions)

constructs

Parameters:

dimensions - amount of dimenstions

NormalDistributionCluster

public NormalDistributionCluster(double[] mx,
                                 double[] var)

convenience constructor to create already trained distribution cluster

Parameters:

mx - precooked expectation values

var - precooked variance

Method Detail

getVar

public double[] getVar()

lazily calculate and return variance cluster

Returns:

variance cluster

train

public void train(double[] samples)

perform sample image training - cumulate values and compute moments

Specified by:

train in interface Cluster

Overrides:

train in class AbstractBaseCluster

Parameters:

samples -

getQuads

public double[] getQuads()

setQuads

public void setQuads(double[] quads)

setVar

public void setVar(double[] var)