[R] AR1 model using ARIMA

[T.Riedle]

Dear R users,
I am trying to run a rolling window AR1 model by combining the rollapply() with the arima() function. Hence, my code looks as follows:

rollingarma<-rollapply(data,width=36,function(data) coef(arima(data,order=c(1,0,0))))
Error in arima(data, order = c(1, 0, 0)) :
  non-stationary AR part from CSS

However, what is wrong with my code?

Can I use the arma() function as alternative? In this case the code is


rollingarma<-rollapply(data,width=36,function(data) coef(arma(data,order=c(1,0),include.intercept = FALSE)))

There were 50 or more warnings (use warnings() to see the first 50)

> warnings()

Warning messages:

1: In optim(coef, err, gr = NULL, hessian = TRUE, ...) :

  one-dimensional optimization by Nelder-Mead is unreliable:

use "Brent" or optimize() directly

In this case, I get a warning message. How can I use Brent or optimize in this code?

Thanks for your support.

	[[alternative HTML version deleted]]