tech.nmfin.meanreversion.elliott2005

Class Elliott2005DLM

java.lang.Object

dev.nm.stat.dlm.univariate.DLM

tech.nmfin.meanreversion.elliott2005.Elliott2005DLM

Direct Known Subclasses:

ElliottOnlineFilter

public class Elliott2005DLM
extends DLM

This class implements the Kalman filter model as in Elliott's paper. It is a simplified version of the general univariate Kalman model. Specifically, the controls in the state equation are 1's; the state coefficient in the observation equation is 1.

See Also:

"Elliott, van der Hoek, and Malcolm, "Pairs trading," Quantitative Finance, 2005."

Constructor Summary

Constructors

Constructor and Description
Elliott2005DLM(double x0, double A, double B, double C, double D) Constructs an Elliott's Kalman filter model.
Elliott2005DLM(double x0, double A, double B, double C, double D, RandomStandardNormalGenerator rnorm) Constructs an Elliott's Kalman filter model.
Elliott2005DLM(Elliott2005DLM that) Copy constructor.

Method Summary

Modifier and Type	Method and Description
double	A() Gets A as in eq. 18.
double	B() Gets B as in eq. 18.
double	C() Gets C as in eq. 18.
double	D() Gets D as in eq. 19, the variance.
double	holdingTime(double c) Gets a suggested holding time based on an OU process.
double	holdingTimeByThreshold(double threshold) Gets a suggested holding time based on an OU process.
double	mu() Gets the long term mean.
double[]	simObs(int T) Simulates a sequence of observations per the model specification.
double[]	simStates(int T) Simulates a sequence of states per the model specification.
double	threshold(double c) Gets a suggested trading threshold based on an OU process.

Methods inherited from class dev.nm.stat.dlm.univariate.DLM

C0, getObsDimension, getObservationModel, getStateDimension, getStateModel, m0

Methods inherited from class java.lang.Object

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

Constructor Detail

Elliott2005DLM

public Elliott2005DLM(double x0,
                      double A,
                      double B,
                      double C,
                      double D,
                      RandomStandardNormalGenerator rnorm)

Constructs an Elliott's Kalman filter model.

Parameters:

x0 - mean for the initial state

A - A as in eq. 18, A >= 0

B - B as in eq. 18, 0 < B < 1

C - C as in eq. 18, C > 0

D - D as in eq. 19, D > 0

rnorm - a standard Gaussian random number generator

Elliott2005DLM

public Elliott2005DLM(double x0,
                      double A,
                      double B,
                      double C,
                      double D)

Constructs an Elliott's Kalman filter model.

Parameters:

x0 - mean for the initial state

A - A as in eq. 18

B - B as in eq. 18

C - C as in eq. 18

D - D as in eq. 19

Elliott2005DLM

public Elliott2005DLM(Elliott2005DLM that)

Copy constructor.

Parameters:

that - another Elliott's Kalman filter model

Method Detail

A

public double A()

Gets A as in eq. 18.

Returns:

A

B

public double B()

Gets B as in eq. 18.

Returns:

B

C

public double C()

Gets C as in eq. 18.

Returns:

C

D

public double D()

Gets D as in eq. 19, the variance.

Returns:

D

simStates

public double[] simStates(int T)

Simulates a sequence of states per the model specification.

Parameters:

T - the length of the sequence

Returns:

a sequence of simulated states

simObs

public double[] simObs(int T)

Simulates a sequence of observations per the model specification.

Parameters:

T - the length of the sequence

Returns:

a sequence of simulated observations

mu

public double mu()

Gets the long term mean.

Returns:

the long term mean

threshold

public double threshold(double c)

Gets a suggested trading threshold based on an OU process.

Parameters:

c - c as in eq. 16

Returns:

the trading threshold

holdingTime

public double holdingTime(double c)

Gets a suggested holding time based on an OU process.

Parameters:

c - c as in eq. 16

Returns:

the holding time

holdingTimeByThreshold

public double holdingTimeByThreshold(double threshold)

Gets a suggested holding time based on an OU process.

Parameters:

threshold - a trading threshold

Returns:

the holding time