fixConstants: Fix TSmodel Coefficients (Parameters) to Constants

Description

Fix specified coefficients to constant values or any coefficients within fuzz of 0.0 or 1.0 to exactly 0.0 or 1.0. This will not change the model much but will affect some estimation techniques and information criteria results, as these are considered to be constants rather than coefficients.

Usage

fixConstants(model, fuzz=1e-5, constants=NULL)

Arguments

model

an object of class TSmodel.

fuzz

absolute difference to be considered equivalent.

constants

NULL or a list of logical arrays.

Value

An object of class 'SS' 'TSmodel' with some array entries set to constants 0.0 or 1.0.

Details

If constants is not NULL then parameters within fuzz of 0.0 or 1.0 are set as constants 0.0 or 1.0. If constants is not NULL then it should be a list with logical arrays named F, G ..., with TRUE corresponding to any array elements which are to be treated as constant.

Examples

Run this code

# NOT RUN {
    f <- array(c(.5,.3,.2,.4),c(2,2))
    h <- array(c(1,0,0,1),c(2,2))
    k <- array(c(.5,.3,.2,.4),c(2,2))
    ss <- SS(F=f,G=NULL,H=h,K=k)
    ss
    coef(ss)
    ss <- fixConstants(ss, constants=list(
              F = matrix(c(TRUE, FALSE, FALSE, FALSE), 2,2)))
    ss
    coef(ss)
    data("eg1.DSE.data.diff", package="dse")
    model <- toARMA(toSS(estVARXls(eg1.DSE.data.diff)))
    model <- fixConstants(model)
# }