org.hawkular.datamining.forecast.stats

Class AugmentedDickeyFullerTest

java.lang.Object

org.hawkular.datamining.forecast.stats.AugmentedDickeyFullerTest

public class AugmentedDickeyFullerTest
extends Object

Augmented Dickey-Fuller test

Null hypothesis H0: Time series contains unit root (is non stationary)

Implemented variants:

• c - with constant (should be used for testing trend stationary)
• ct - with constant ant time trend
• nc - no constant

Equations for ADF test

• yDiff = alpha + beta*t + gamma*yLag + sum(deltaLagT*yLagDiff)
• dfStatistics = gammaHat/standardError(gammaHat)

Statistical tables adapted from: Python statsmodels

Author:

Pavol Loffay

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	AugmentedDickeyFullerTest.Type ADF test variants

Constructor Summary

Constructors

Constructor and Description
AugmentedDickeyFullerTest(double[] x, int maxLag) Default constructor for testing trend stationary time series.
AugmentedDickeyFullerTest(double[] x, int maxLag, AugmentedDickeyFullerTest.Type type)

Method Summary

Modifier and Type	Method and Description
AugmentedDickeyFullerTest.Type	getType()
double	pValue()
double	statistics()

Methods inherited from class java.lang.Object

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

Constructor Detail

AugmentedDickeyFullerTest

public AugmentedDickeyFullerTest(double[] x,
                                 int maxLag)

Default constructor for testing trend stationary time series. Type is set to InterceptNoTimeTrend (in R 'c' - constant)

Parameters:

x -

maxLag -

AugmentedDickeyFullerTest

public AugmentedDickeyFullerTest(double[] x,
                                 int maxLag,
                                 AugmentedDickeyFullerTest.Type type)

Method Detail

getType

public AugmentedDickeyFullerTest.Type getType()

pValue

public double pValue()

statistics

public double statistics()