toSSChol: Convert to Non-Innovation State Space Model

In dse: Dynamic Systems Estimation (Time Series Package)

Description

This function may not be working properly.

Convert to a non-innovations state space representation using the given matrix (Om) as the measurement noise covariance. Om would typically be an estimate of the output noise, such as returned in $estimates$cov of the function l (l.SS or l.ARMA). This assumes that the noise processes in the arbitrary SS representation are white and uncorrelated.

Usage

    toSSChol(model, ...)
    ## S3 method for class 'TSmodel'
toSSChol(model, Om=diag(1,nseriesOutput(model)), ...)
    ## S3 method for class 'TSestModel'
toSSChol(model, Om=NULL, ...)

Arguments

model	An object of class TSmodel.
Om	a matrix to be used as the measurement noise covariance. If Om is not supplied and model is of class TSestModel then model$estimates$cov is used. Otherwise, Om is set to the identity matrix.
...	arguments to be passed to other methods.

Details

Convert to a non-innovations SS representation using a Cholesky decomposition of Om as the coefficient matrix of the output noise.

Value

An object of class 'SS' 'TSmodel' containing a state space model which is not in innovations form.

See Also

toSSinnov

Examples

    data("eg1.DSE.data.diff", package="dse")
    model <- estVARXls(eg1.DSE.data.diff)
    model <- toSSChol(model)

dse documentation built on March 26, 2020, 7:12 p.m.