[R] arima/fixed

[Rolf Turner]

On Fri, 29 Jan 2021 12:47:25 +0000
Nasia Petsa <petsa.athanasia using hotmail.gr> wrote:

> Dear all,
> 
> I have the following problem with determining the argument fixed in
> arima function. What is the length  of argument fixed for an
> ARIMA(1,0,0)(0,1,1) and what is the correct  order for the parameters

Hmm.  The help is indeed pretty opaque, isn't it?  I tried
the following:

set.seed(42)
x <- rnorm(300)
f1 <- arima(x,order=c(1,0,0),seasonal=list(order=c(0,1,1),period=6))
coef(f1) # Which gave:
>        ar1       sma1 
> -0.0169276 -0.9999999 

f2 <- arima(x,order=c(1,0,0),seasonal=list(order=c(0,1,1),period=6),
            fixed=c(-0.5,NA),transform.pars=FALSE)
coef(f2) # Which gave:
> ar1 sma1 
> -0.5 -1.0 

So this looks like it's doing the right thing .... ???

Printing f2 indicates that the standard error of "ar1" is 0, which
makes sense since it's fixed.

I hope this helps.  Perhaps someone who actually knows what they are
talking about will chime in.

cheers,

Rolf Turner

-- 
Honorary Research Fellow
Department of Statistics
University of Auckland
Phone: +64-9-373-7599 ext. 88276