Package org.deidentifier.arx.framework.check.distribution

Klasse DistributionAggregateFunction

java.lang.Object

org.deidentifier.arx.framework.check.distribution.DistributionAggregateFunction

Alle implementierten Schnittstellen:

Serializable

Bekannte direkte Unterklassen:

DistributionAggregateFunction.DistributionAggregateFunctionArithmeticMean, DistributionAggregateFunction.DistributionAggregateFunctionGeometricMean, DistributionAggregateFunction.DistributionAggregateFunctionInterval, DistributionAggregateFunction.DistributionAggregateFunctionMedian, DistributionAggregateFunction.DistributionAggregateFunctionMode, DistributionAggregateFunction.DistributionAggregateFunctionModeWithDistributionFallback, DistributionAggregateFunction.DistributionAggregateFunctionSet

public abstract class DistributionAggregateFunction
extends Object
implements Serializable

This abstract class represents a function that aggregates values from a frequency distribution

Siehe auch:

• Serialisierte Form

Verschachtelte Klassen - Übersicht

Verschachtelte Klassen

Modifizierer und Typ	Klasse	Beschreibung
static class	DistributionAggregateFunction.DistributionAggregateFunctionArithmeticMean	This class calculates the arithmetic mean for a given distribution.
static class	DistributionAggregateFunction.DistributionAggregateFunctionGeometricMean	This class calculates the geometric mean for a given distribution.
static class	DistributionAggregateFunction.DistributionAggregateFunctionInterval	This class calculates the mode for a given distribution.
static class	DistributionAggregateFunction.DistributionAggregateFunctionMedian	This class calculates the median for a given distribution.
static class	DistributionAggregateFunction.DistributionAggregateFunctionMode	This class calculates the mode for a given distribution.
static class	DistributionAggregateFunction.DistributionAggregateFunctionModeWithDistributionFallback	This class calculates the mode for a given distribution falling back to drawing from multiple values that would qualify as mode using the provided distribution
static class	DistributionAggregateFunction.DistributionAggregateFunctionSet	This class calculates a set for a given distribution.

Feldübersicht

Felder

Modifizierer und Typ	Feld	Beschreibung
protected String[]	dictionary	Dictionary
protected boolean	ignoreMissingData	Whether or not null values should be ignored
protected DataType<?>	type	Type

Konstruktorübersicht

Konstruktoren

Konstruktor	Beschreibung
DistributionAggregateFunction(boolean ignoreMissingData, boolean typePreserving)	Instantiates a new function.

Methodenübersicht

Modifizierer und Typ	Methode	Beschreibung
protected <T> void	addAll(org.apache.commons.math3.stat.descriptive.DescriptiveStatistics statistics, Distribution distribution, DataType.DataTypeWithRatioScale<T> type, double offset)	Adds all values from the distribution to the given descriptive statistics object
abstract <T> String	aggregate(Distribution distribution)	This function returns an aggregate value.
abstract DistributionAggregateFunction	clone()	Clones this function
abstract <T> double	getError(Distribution distribution)	Returns the normalized error induced by aggregation.
<T> double	getInformationLoss(Distribution distribution)	This will return the normalized number of aggregated values in range [1/#distinct-values, 1].
protected <T> double[]	getMinMax(String[] dictionary, DataType.DataTypeWithRatioScale<T> type)	Returns the minimum and maximum value
protected double	getNMSE(double min, double max, double[] values, double aggregate)	Calculates the mean square error after normalizing everything into [0,1]
void	initialize(String[] dictionary, DataType<?> type)	Initializes the function
boolean	isIgnoreMissingData()	Returns whether this function ignores missing data
boolean	isTypePreserving()	Returns whether this is a type-preserving function

Von Klasse geerbte Methoden java.lang.Object

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

Felddetails

ignoreMissingData

protected boolean ignoreMissingData

Whether or not null values should be ignored

dictionary

protected transient String[] dictionary

Dictionary

type

protected transient DataType<?> type

Type

Konstruktordetails

DistributionAggregateFunction

public DistributionAggregateFunction(boolean ignoreMissingData,
 boolean typePreserving)

Instantiates a new function.

Parameter:

ignoreMissingData -

typePreserving -

Methodendetails

aggregate

public abstract <T> String aggregate(Distribution distribution)

This function returns an aggregate value.

Parameter:

distribution -

dictionary -

type -

Gibt zurück:

the string

clone

public abstract DistributionAggregateFunction clone()

Clones this function

Setzt außer Kraft:

clone in Klasse Object

getError

public abstract <T> double getError(Distribution distribution)

Returns the normalized error induced by aggregation. In most cases this will be the mean squared error normalized into [0,1]. In case of generalization, it will return the normalized generalization level (also called generalization intensity). In case of intervals, it will return the normalized number of aggregated values.

Parameter:

distribution -

Gibt zurück:

getInformationLoss

public <T> double getInformationLoss(Distribution distribution)

This will return the normalized number of aggregated values in range [1/#distinct-values, 1].

Parameter:

distribution -

Gibt zurück:

initialize

public void initialize(String[] dictionary,
 DataType<?> type)

Initializes the function

Parameter:

dictionary -

type -

hierarchy -

isTypePreserving

public boolean isTypePreserving()

Returns whether this is a type-preserving function

Gibt zurück:

isIgnoreMissingData

public boolean isIgnoreMissingData()

Returns whether this function ignores missing data

Gibt zurück:

addAll

protected <T> void addAll(org.apache.commons.math3.stat.descriptive.DescriptiveStatistics statistics,
 Distribution distribution,
 DataType.DataTypeWithRatioScale<T> type,
 double offset)

Adds all values from the distribution to the given descriptive statistics object

Parameter:

statistics -

distribution -

type -

offset - will be added to values

getMinMax

protected <T> double[] getMinMax(String[] dictionary,
 DataType.DataTypeWithRatioScale<T> type)

Returns the minimum and maximum value

Parameter:

dictionary -

type -

Gibt zurück:

getNMSE

protected double getNMSE(double min,
 double max,
 double[] values,
 double aggregate)

Calculates the mean square error after normalizing everything into [0,1]

Parameter:

min -

max -

values -

aggregate -

Gibt zurück: