CODEWITHSUNDEEP
rarima
Candice2020
Candice2020

Reputation: 31

Unused arguments of sarima()

When I am trying to run 

sarima(log(AirPassengers),0,1,1,0,1,1,12)

It returns

Error in sarima(log(AirPassengers), 0, 1, 1, 0, 1, 1, 12) : unused arguments (1, 1, 12)

The dataset is under the package of astsa package. The data type is ts type.

class(AirPassengers)
[1] "ts"

I don't know the reason of this error. I ran the same code in DataCamp, it works.

Upvotes: 0

Views: 1266

Answers (1)

Maurits Evers
Maurits Evers

Reputation: 50718

A few general comments:

  1. You can fit SARIMA models with base R's arima function; there is really no need to load an additional package for fitting basic SARIMA models.
  2. The AirPassengers dataset is part of base R, not part of astsa.

In arima, you specify the non-seasonal and seasonal parameters through function arguments order and seasonal, respectively. You can find detailed information what arguments the function takes if you type ?arima into an R terminal.

For example, to specify a SARIMA(0,1,1)(0,1,1)12 model you do

arima(log(AirPassengers), order = c(0,1,1), seasonal = list(order = c(0,1,1), period = 12))

Note that the seasonal argument is a list with the two elements order (denoting the order of seasonal AR terms, order of seasonal differencing, and order of seasonal MA terms) and period.

The error you're seeing with astsa::sarima seems to stem from too many arguments; check that you really have 3 non-seasonal arguments, 3 (optional) seasonal arguments, and one (optional) period.

Above model would be

sarima(log(AirPassengers), 0, 1, 1, 0, 1, 1, 12)

which produces the same fit results as arima (in fact sarima just calls arima "under the hood").

Upvotes: 1

Related Questions