com.numericalmethod.suanshu.stats.dlm.multivariate

Class DLMSim

java.lang.Object

com.numericalmethod.suanshu.stats.dlm.multivariate.DLMSim

All Implemented Interfaces:

java.io.Serializable

public class DLMSim
extends java.lang.Object
implements java.io.Serializable

This is a simulator for a multivariate controlled dynamic linear model process. For (t ≥ 1), a controlled DLM takes the following form:

Observation Equation:

y_t = F_t * x_t + v_t,

State Equation:

x_t = G_t * x_t-1 + H_t * u_t + w_t,

Given the model parameters, the time series of control variables {u_t} and an integer (length) T, This simulator generates both the states {x_t} and observations {y_t}.

See Also:

"G. Petris et al., "Dynamic Linear Models with R," New York, Springer, 2009, ch. 2, pp. 31-84.", Serialized Form

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	DLMSim.Innovation a simulated innovation

Constructor Summary

Constructors

Constructor and Description
DLMSim(DLM model, NormalRvg rmvnorm) Simulate a multivariate controlled dynamic linear model process.

Method Summary

Modifier and Type	Method and Description
DLMSim.Innovation	next(Vector u) Get the next innovation.

Methods inherited from class java.lang.Object

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

Constructor Detail

DLMSim

public DLMSim(DLM model,
              NormalRvg rmvnorm)

Simulate a multivariate controlled dynamic linear model process.

Parameters:

model - a multivariate controlled dynamic linear model

rmvnorm - a standard multivariate Gaussian random vector generator (for seeding)

Method Detail

next

public DLMSim.Innovation next(Vector u)

Get the next innovation.

Parameters:

u - the control vector as of time t - 1; used not for the first innovation (initial state)

Returns:

a simulated innovation