dev.nm.stat.timeseries.linear.multivariate.arima

Class VARIMAModel

java.lang.Object

dev.nm.stat.timeseries.linear.multivariate.arima.VARIMAXModel

dev.nm.stat.timeseries.linear.multivariate.arima.VARIMAModel

Direct Known Subclasses:

VARMAModel

public class VARIMAModel
extends VARIMAXModel

An ARIMA(p, d, q) process, \(Y_{t}\), is such that \[ X_t = (1 - L)^d Y_t \] where L is the lag operator, d the order of difference, \(X_{t}\) an ARMA(p, q) process, for which \[ X_t = \mu + \Sigma \phi_i X_{t-i} + \Sigma \theta_j \epsilon_{t-j} + \epsilon_t, \] \(X_{t}\), μ and \(\epsilon _{t}\) are n-dimensional vectors. The (n * n) matrices \({\phi_i}\) and \({\theta_j}\) are the AR and MA coefficients respectively.

See Also:

Wikipedia: Autoregressive integrated moving average

Constructor Summary

Constructors

Constructor and Description
VARIMAModel(ARIMAModel model) Construct a multivariate model from a univariate ARIMA model.
VARIMAModel(Matrix[] phi, int d, Matrix[] theta) Construct a multivariate ARIMA model with unit variance and zero-intercept (mu).
VARIMAModel(Matrix[] phi, int d, Matrix[] theta, Matrix sigma) Construct a multivariate ARIMA model with zero-intercept (mu).
VARIMAModel(VARIMAModel that) Copy constructor.
VARIMAModel(Vector mu, Matrix[] phi, int d, Matrix[] theta) Construct a multivariate ARIMA model with unit variance.
VARIMAModel(Vector mu, Matrix[] phi, int d, Matrix[] theta, Matrix sigma) Construct a multivariate ARIMA model.

Method Summary

Modifier and Type	Method and Description
VARMAModel	getVARMA() Get the ARMA part of this ARIMA model, essentially ignoring the differencing.

Methods inherited from class dev.nm.stat.timeseries.linear.multivariate.arima.VARIMAXModel

AR, d, dimension, getVARMAX, MA, maxPQ, mu, p, phi, psi, q, sigma, theta

Methods inherited from class java.lang.Object

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

Constructor Detail

VARIMAModel

public VARIMAModel(Vector mu,
                   Matrix[] phi,
                   int d,
                   Matrix[] theta,
                   Matrix sigma)

Construct a multivariate ARIMA model.

Parameters:

mu - the intercept (constant) vector

phi - the AR coefficients (excluding the initial 1); null if no AR coefficient

d - the order of integration

theta - the MA coefficients (excluding the initial 1); null if no MA coefficient

sigma - the white noise covariance matrix

VARIMAModel

public VARIMAModel(Vector mu,
                   Matrix[] phi,
                   int d,
                   Matrix[] theta)

Construct a multivariate ARIMA model with unit variance.

Parameters:

mu - the intercept (constant) vector

phi - the AR coefficients (excluding the initial 1); null if no AR coefficient

d - the order of integration

theta - the MA coefficients (excluding the initial 1); null if no MA coefficient

VARIMAModel

public VARIMAModel(Matrix[] phi,
                   int d,
                   Matrix[] theta,
                   Matrix sigma)

Construct a multivariate ARIMA model with zero-intercept (mu).

Parameters:

phi - the AR coefficients (excluding the initial 1); null if no AR coefficient

d - the order of integration

theta - the MA coefficients (excluding the initial 1); null if no MA coefficient

sigma - the white noise covariance matrix

VARIMAModel

public VARIMAModel(Matrix[] phi,
                   int d,
                   Matrix[] theta)

Construct a multivariate ARIMA model with unit variance and zero-intercept (mu).

Parameters:

phi - the AR coefficients (excluding the initial 1); null if no AR coefficient

d - the order of integration

theta - the MA coefficients (excluding the initial 1); null if no MA coefficient

VARIMAModel

public VARIMAModel(ARIMAModel model)

Construct a multivariate model from a univariate ARIMA model.

Parameters:

model - a univariate ARIMA model

VARIMAModel

public VARIMAModel(VARIMAModel that)

Copy constructor.

Parameters:

that - a multivariate ARIMA model

Method Detail

getVARMA

public VARMAModel getVARMA()

Get the ARMA part of this ARIMA model, essentially ignoring the differencing.

Returns:

the ARMA part