toSS: Convert to State Space Model

In dse: Dynamic Systems Estimation (Time Series Package)

Description

Convert a model to state space form.

Usage

    toSS(model, ...)
    ## S3 method for class 'ARMA'
toSS(model, ...)
    ## S3 method for class 'SS'
toSS(model, ...)
    ## S3 method for class 'TSestModel'
toSS(model, ...)
    
    toSSaugment(model, ...)
    ## S3 method for class 'ARMA'
toSSaugment(model, fuzz=1e-14, ...)
    ## S3 method for class 'TSestModel'
toSSaugment(model, ...)
    
    toSSnested(model, ...)
    ## S3 method for class 'ARMA'
toSSnested(model, n=NULL, Aoki=FALSE, ...)
    ## S3 method for class 'SS'
toSSnested(model, n=NULL, Aoki=FALSE, ...)
    ## S3 method for class 'TSestModel'
toSSnested(model, ...)
    
    

Arguments

model	An object of class TSmodel.
n	If n is specified then it is used as the state dimension when the markov parameter conversion technique is required.
Aoki	logical indicating if Aoki's method (which does not work in general) should be tried.
fuzz	if the zero lag term of polynomials A and B are within fuzz of the identitity matrix then they are not inverted. (i.e. they are assumed to be identity.)
...	arguments to be passed to other methods.

Details

If the order of the AR polynomial equals or exceeds the MA polynomial (and the input polynomial) then the model is converted by state augmentation. Otherwise, it is converted by approximating the markov coefficients a la Mittnik. (This may not always work very well. Compare the results to check.)

Value

A state space model in an object of class 'SS' 'TSmodel'.

Examples

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

dse documentation built on March 4, 2020, 3:01 a.m.