org.apache.commons.math3.stat.descriptive.rank

Class PSquarePercentile

java.lang.Object

org.apache.commons.math3.stat.descriptive.AbstractUnivariateStatistic

org.apache.commons.math3.stat.descriptive.AbstractStorelessUnivariateStatistic

org.apache.commons.math3.stat.descriptive.rank.PSquarePercentile

All Implemented Interfaces:

java.io.Serializable, StorelessUnivariateStatistic, UnivariateStatistic, MathArrays.Function

public class PSquarePercentile
extends AbstractStorelessUnivariateStatistic
implements StorelessUnivariateStatistic, java.io.Serializable

A StorelessUnivariateStatistic estimating percentiles using the P² Algorithm as explained by Raj Jain and Imrich Chlamtac in P² Algorithm for Dynamic Calculation of Quantiles and Histogram Without Storing Observations.

Note: This implementation is not synchronized and produces an approximate result. For small samples, where data can be stored and processed in memory, Percentile should be used.

See Also:

Serialized Form

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
private static class	PSquarePercentile.FixedCapacityList<E> A simple fixed capacity list that has an upper bound to growth.
private static class	PSquarePercentile.Marker The class modeling the attributes of the marker of the P-square algorithm
private static class	PSquarePercentile.Markers Markers is an encapsulation of the five markers/buckets as indicated in the original works.
protected static interface	PSquarePercentile.PSquareMarkers An interface that encapsulates abstractions of the P-square algorithm markers as is explained in the original works.

Field Summary

Fields

Modifier and Type	Field and Description
private long	countOfObservations Counter to count the values/observations accepted into this data set
private static java.text.DecimalFormat	DECIMAL_FORMAT A decimal formatter for print convenience
private static double	DEFAULT_QUANTILE_DESIRED A Default quantile needed in case if user prefers to use default no argument constructor.
private java.util.List<java.lang.Double>	initialFive Initial list of 5 numbers corresponding to 5 markers.
private double	lastObservation lastObservation is the last observation value/input sample.
private PSquarePercentile.PSquareMarkers	markers Markers is the marker collection object which comes to effect only after 5 values are inserted
private static int	PSQUARE_CONSTANT The maximum array size used for psquare algorithm
private double	pValue Computed p value (i,e percentile value of data set hither to received)
private double	quantile The quantile needed should be in range of 0-1.
private static long	serialVersionUID Serial ID

Constructor Summary

Constructors

Constructor and Description
PSquarePercentile() Default constructor that assumes a default quantile needed
PSquarePercentile(double p) Constructs a PSquarePercentile with the specific percentile value.

Method Summary

Methods

Modifier and Type	Method and Description
void	clear() Clears the internal state of the Statistic
StorelessUnivariateStatistic	copy() Returns a copy of the statistic with the same internal state.
boolean	equals(java.lang.Object o) Returns true iff o is a PSquarePercentile returning the same values as this for getResult() and getN() and also having equal markers
long	getN() Returns the number of values that have been added.
double	getResult() Returns the current value of the Statistic.
int	hashCode() Returns hash code based on getResult() and getN()
void	increment(double observation) Updates the internal state of the statistic to reflect the addition of the new value.
private double	maximum()
private double	minimum()
static PSquarePercentile.PSquareMarkers	newMarkers(java.util.List<java.lang.Double> initialFive, double p) A creation method to build Markers
double	quantile() Returns the quantile estimated by this statistic in the range [0.0-1.0]
java.lang.String	toString() Returns a string containing the last observation, the current estimate of the quantile and all markers.

Methods inherited from class org.apache.commons.math3.stat.descriptive.AbstractStorelessUnivariateStatistic

evaluate, evaluate, incrementAll, incrementAll

Methods inherited from class org.apache.commons.math3.stat.descriptive.AbstractUnivariateStatistic

evaluate, getData, getDataRef, setData, setData, test, test, test, test

Methods inherited from class java.lang.Object

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

Methods inherited from interface org.apache.commons.math3.stat.descriptive.StorelessUnivariateStatistic

incrementAll, incrementAll

Methods inherited from interface org.apache.commons.math3.stat.descriptive.UnivariateStatistic

evaluate, evaluate

Field Detail

PSQUARE_CONSTANT

private static final int PSQUARE_CONSTANT

The maximum array size used for psquare algorithm

See Also:

Constant Field Values

DEFAULT_QUANTILE_DESIRED

private static final double DEFAULT_QUANTILE_DESIRED

A Default quantile needed in case if user prefers to use default no argument constructor.

See Also:

Constant Field Values

serialVersionUID

private static final long serialVersionUID

Serial ID

See Also:

Constant Field Values

DECIMAL_FORMAT

private static final java.text.DecimalFormat DECIMAL_FORMAT

A decimal formatter for print convenience

initialFive

private final java.util.List<java.lang.Double> initialFive

Initial list of 5 numbers corresponding to 5 markers. NOTE:watch out for the add methods that are overloaded

quantile

private final double quantile

The quantile needed should be in range of 0-1. The constructor PSquarePercentile(double) ensures that passed in percentile is divided by 100.

lastObservation

private transient double lastObservation

lastObservation is the last observation value/input sample. No need to serialize

markers

private PSquarePercentile.PSquareMarkers markers

Markers is the marker collection object which comes to effect only after 5 values are inserted

pValue

private double pValue

Computed p value (i,e percentile value of data set hither to received)

countOfObservations

private long countOfObservations

Counter to count the values/observations accepted into this data set

Constructor Detail

PSquarePercentile

public PSquarePercentile(double p)

Constructs a PSquarePercentile with the specific percentile value.

Parameters:

p - the percentile

Throws:

OutOfRangeException - if p is not greater than 0 and less than or equal to 100

PSquarePercentile

PSquarePercentile()

Default constructor that assumes a default quantile needed

Method Detail

hashCode

public int hashCode()

Returns hash code based on getResult() and getN()

Overrides:

hashCode in class AbstractStorelessUnivariateStatistic

Returns:

hash code

equals

public boolean equals(java.lang.Object o)

Returns true iff o is a PSquarePercentile returning the same values as this for getResult() and getN() and also having equal markers

Overrides:

equals in class AbstractStorelessUnivariateStatistic

Parameters:

o - object to compare

Returns:

true if o is a PSquarePercentile with equivalent internal state

increment

public void increment(double observation)

Updates the internal state of the statistic to reflect the addition of the new value.The internal state updated due to the new value in this context is basically of the marker positions and computation of the approximate quantile.

Specified by:

increment in interface StorelessUnivariateStatistic

Specified by:

increment in class AbstractStorelessUnivariateStatistic

Parameters:

observation - the observation currently being added.

toString

public java.lang.String toString()

Returns a string containing the last observation, the current estimate of the quantile and all markers.

Overrides:

toString in class java.lang.Object

Returns:

string representation of state data

getN

public long getN()

Returns the number of values that have been added.

Specified by:

getN in interface StorelessUnivariateStatistic

Returns:

the number of values.

copy

public StorelessUnivariateStatistic copy()

Returns a copy of the statistic with the same internal state.

Specified by:

copy in interface StorelessUnivariateStatistic

Specified by:

copy in interface UnivariateStatistic

Specified by:

copy in class AbstractStorelessUnivariateStatistic

Returns:

a copy of the statistic

quantile

public double quantile()

Returns the quantile estimated by this statistic in the range [0.0-1.0]

Returns:

quantile estimated by getResult()

clear

public void clear()

Clears the internal state of the Statistic. This basically clears all the markers, the initialFive list and sets countOfObservations to 0.

Specified by:

clear in interface StorelessUnivariateStatistic

Specified by:

clear in class AbstractStorelessUnivariateStatistic

getResult

public double getResult()

Returns the current value of the Statistic.

Specified by:

getResult in interface StorelessUnivariateStatistic

Specified by:

getResult in class AbstractStorelessUnivariateStatistic

Returns:

value of the statistic, Double.NaN if it has been cleared or just instantiated.

maximum

private double maximum()

Returns:

maximum in the data set added to this statistic

minimum

private double minimum()

Returns:

minimum in the data set added to this statistic

newMarkers

public static PSquarePercentile.PSquareMarkers newMarkers(java.util.List<java.lang.Double> initialFive,
                                          double p)

A creation method to build Markers

Parameters:

initialFive - list of initial five elements

p - the quantile desired

Returns:

an instance of PSquareMarkers