com.datumbox.framework.core.statistics.timeseries

Class Smoothing

java.lang.Object

com.datumbox.framework.core.statistics.timeseries.Smoothing

public class Smoothing
extends java.lang.Object

This class provides several Smoothing and moving average methods.

Constructor Summary

Constructors

Constructor and Description
Smoothing()

Method Summary

Modifier and Type	Method and Description
static double	holtWintersSmoothing(FlatDataList flatDataList, int season_length, double alpha, double beta, double gamma, double dev_gamma) Smooth supplied timeline internalData 3 ways - overall, by trend and by season.
static double	simpleExponentialSmoothing(FlatDataList flatDataList, double a) Simple Explonential Smoothing
static double	simpleExponentialSmoothingQuick(double Ytminus1, double Stminus1, double a) Simple Explonential Smoothing: Calculates Ft+1 by using previous results.
static double	simpleMovingAverage(FlatDataList flatDataList, int N) Simple Moving Average
static double	simpleMovingAverageQuick(double Yt, double YtminusN, double Ft, int N) Simple Moving Average: Calculates Ft+1 by using previous results.
static double	weightedMovingAverage(FlatDataList flatDataList, int N) Weighted Moving Average

Methods inherited from class java.lang.Object

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

Constructor Detail

Smoothing

public Smoothing()

Method Detail

simpleMovingAverage

public static double simpleMovingAverage(FlatDataList flatDataList,
                                         int N)

Simple Moving Average

Parameters:

flatDataList -

N -

Returns:

simpleMovingAverageQuick

public static double simpleMovingAverageQuick(double Yt,
                                              double YtminusN,
                                              double Ft,
                                              int N)

Simple Moving Average: Calculates Ft+1 by using previous results. A quick version of the simpleMovingAverage

Parameters:

Yt -

YtminusN -

Ft -

N -

Returns:

weightedMovingAverage

public static double weightedMovingAverage(FlatDataList flatDataList,
                                           int N)

Weighted Moving Average

Parameters:

flatDataList -

N -

Returns:

simpleExponentialSmoothing

public static double simpleExponentialSmoothing(FlatDataList flatDataList,
                                                double a)

Simple Explonential Smoothing

Parameters:

flatDataList -

a -

Returns:

simpleExponentialSmoothingQuick

public static double simpleExponentialSmoothingQuick(double Ytminus1,
                                                     double Stminus1,
                                                     double a)

Simple Explonential Smoothing: Calculates Ft+1 by using previous results. A quick version of the simpleExponentialSmoothing

Parameters:

Ytminus1 -

Stminus1 -

a -

Returns:

holtWintersSmoothing

public static double holtWintersSmoothing(FlatDataList flatDataList,
                                          int season_length,
                                          double alpha,
                                          double beta,
                                          double gamma,
                                          double dev_gamma)

Smooth supplied timeline internalData 3 ways - overall, by trend and by season. Ported from https://github.com/ianbarber/PHPIR/blob/master/holtwinters.php

Parameters:

flatDataList - - list of internalData

season_length - - the number of entries that represent a 'season'. example = 7

alpha - - internalData smoothing factor. example = 0.2

beta - - trend smoothing factor. example = 0.01

gamma - - seasonality smoothing factor. example = 0.01

dev_gamma - - smoothing factor for deviations. example = 0.1

Returns: