I need to estimate parameters of an AR model which is in the form of AR(1,11) it means that coefficients of AR orders from order 2 until order 10 are zero. How can I estimate these two parameters in R since arima function only accepts p as the order of the AR component. Note that this model has a different structure than Seasonal AR.



Wouldn't this be possible by using the FIXED option in the ARIMA syntax? In this case,

output <- arima(x, order=c(11,0,0), fixed = c(NA,0,0,0,0,0,0,0,0,0,NA))

The order statement supplies the number of lags for the AR process, and the fixed vector is "NA" for any lags you want to estimate and zero otherwise.

  • $\begingroup$ thank you, I didn't know that we can assign "NA" for parameters that we need. I will try this. $\endgroup$ – Fred Apr 24 '13 at 21:15
  • $\begingroup$ Sure thing, let me know if that resolves your problem. $\endgroup$ – Corcovado Apr 24 '13 at 21:21
  • $\begingroup$ Yes it did but we should add another NA for q order like this output <- arima(x, order=c(11,0,0), fixed = c(NA,0,0,0,0,0,0,0,0,0,NA, NA)) $\endgroup$ – Fred Apr 24 '13 at 21:30

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.