Package com.tdunning.math.stats

Class AbstractTDigest

java.lang.Object

com.tdunning.math.stats.TDigest

com.tdunning.math.stats.AbstractTDigest

All Implemented Interfaces:

java.io.Serializable

Direct Known Subclasses:

AVLTreeDigest, MergingDigest

public abstract class AbstractTDigest
extends TDigest

See Also:

Serialized Form

Field Summary

Fields

Modifier and Type	Field	Description
(package private) java.util.Random	gen
(package private) boolean	recordAllData

Fields inherited from class com.tdunning.math.stats.TDigest

max, min

Constructor Summary

Constructors

Constructor	Description
AbstractTDigest()

Method Summary

Modifier and Type	Method	Description
void	add(double x)	Adds a sample to a histogram.
(package private) abstract void	add(double x, int w, Centroid base)
void	add(TDigest other)	Add all of the centroids of another TDigest to this one.
protected Centroid	createCentroid(double mean, int id)
(package private) static int	decode(java.nio.ByteBuffer buf)
(package private) static void	encode(java.nio.ByteBuffer buf, int n)
(package private) static double	interpolate(double x, double x0, double x1)
boolean	isRecording()
(package private) static double	quantile(double index, double previousIndex, double nextIndex, double previousMean, double nextMean)	Computes an interpolated value of a quantile that is between two centroids.
TDigest	recordAllData()	Sets up so that all centroids will record all data assigned to them.
(package private) static double	weightedAverage(double x1, double w1, double x2, double w2)	Same as weightedAverageSorted(double, double, double, double) but flips the order of the variables if x2 is greater than x1.
private static double	weightedAverageSorted(double x1, double w1, double x2, double w2)	Compute the weighted average between x1 with a weight of w1 and x2 with a weight of w2.

Methods inherited from class com.tdunning.math.stats.TDigest

add, add, asBytes, asSmallBytes, byteSize, cdf, centroidCount, centroids, checkValue, compress, compression, createAvlTreeDigest, createDigest, createMergingDigest, getMax, getMin, quantile, setMinMax, size, smallByteSize

Methods inherited from class java.lang.Object

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

Field Detail

gen

final java.util.Random gen

recordAllData

boolean recordAllData

Constructor Detail

AbstractTDigest

public AbstractTDigest()

Method Detail

weightedAverage

static double weightedAverage(double x1,
                              double w1,
                              double x2,
                              double w2)

Same as weightedAverageSorted(double, double, double, double) but flips the order of the variables if x2 is greater than x1.

weightedAverageSorted

private static double weightedAverageSorted(double x1,
                                            double w1,
                                            double x2,
                                            double w2)

Compute the weighted average between x1 with a weight of w1 and x2 with a weight of w2. This expects x1 to be less than or equal to x2 and is guaranteed to return a number between x1 and x2.

interpolate

static double interpolate(double x,
                          double x0,
                          double x1)

encode

static void encode(java.nio.ByteBuffer buf,
                   int n)

decode

static int decode(java.nio.ByteBuffer buf)

add

abstract void add(double x,
                  int w,
                  Centroid base)

quantile

static double quantile(double index,
                       double previousIndex,
                       double nextIndex,
                       double previousMean,
                       double nextMean)

Computes an interpolated value of a quantile that is between two centroids. Index is the quantile desired multiplied by the total number of samples - 1.

Parameters:

index - Denormalized quantile desired

previousIndex - The denormalized quantile corresponding to the center of the previous centroid.

nextIndex - The denormalized quantile corresponding to the center of the following centroid.

previousMean - The mean of the previous centroid.

nextMean - The mean of the following centroid.

Returns:

The interpolated mean.

recordAllData

public TDigest recordAllData()

Sets up so that all centroids will record all data assigned to them. For testing only, really.

Specified by:

recordAllData in class TDigest

Returns:

This TDigest so that configurations can be done in fluent style.

isRecording

public boolean isRecording()

Specified by:

isRecording in class TDigest

add

public void add(double x)

Adds a sample to a histogram.

Specified by:

add in class TDigest

Parameters:

x - The value to add.

add

public void add(TDigest other)

Description copied from class: TDigest

Add all of the centroids of another TDigest to this one.

Specified by:

add in class TDigest

Parameters:

other - The other TDigest

createCentroid

protected Centroid createCentroid(double mean,
                                  int id)