Class MultivariateDLMSeries

java.lang.Object
dev.nm.stat.dlm.multivariate.MultivariateDLMSeries
All Implemented Interfaces:
TimeSeries<Integer,MultivariateDLMSim.Innovation,MultivariateDLMSeries.Entry>, Iterable<MultivariateDLMSeries.Entry>

public class MultivariateDLMSeries extends Object implements TimeSeries<Integer,MultivariateDLMSim.Innovation,MultivariateDLMSeries.Entry>
This is a simulator for a multivariate controlled dynamic linear model process. For (t ≥ 1), a controlled DLM takes the following form:

Observation Equation:

yt = Ft * xt + vt,
State Equation:
xt = Gt * xt-1 + Ht * ut + wt,
Given the model parameters, the time series of control variables {ut} and an integer (length) T, This simulator generates both the states {xt} and observations {yt}.

This implementation is appropriate for a short time series when the size is known; otherwise use instead DLMSim.

  • Constructor Details

    • MultivariateDLMSeries

      public MultivariateDLMSeries(int T, MultivariateDLM model, MultivariateIntTimeTimeSeries Ut, NormalRVG rmvnorm)
      Simulate a multivariate controlled dynamic linear model process.
      Parameters:
      T - the length of the multivariate time series (states and observations) to generate
      model - a multivariate controlled dynamic linear model
      Ut - an m-dimensional time series of control variables (length = T); use null if no control
      rmvnorm - a standard multivariate Gaussian random vector generator (for seeding)
    • MultivariateDLMSeries

      public MultivariateDLMSeries(int T, MultivariateDLM model, MultivariateIntTimeTimeSeries Ut)
      Simulate a multivariate controlled dynamic linear model process.
      Parameters:
      T - the length of the multivariate time series (states and observations) to generate
      model - a multivariate controlled dynamic linear model
      Ut - an m-dimensional time series of control variables (length = T); use null if no control
    • MultivariateDLMSeries

      public MultivariateDLMSeries(int T, MultivariateDLM model)
      Simulate a multivariate controlled dynamic linear model process.
      Parameters:
      T - the length of the multivariate time series (states and observations) to generate
      model - a multivariate controlled dynamic linear model
  • Method Details