Solved – Time series forecast from STL using ARIMA method for specific order

I am trying to forecast daily sales and have found that for the seasonally adjusted component a non-seasonal ARIMA (3,1,6) and a seasonal naive for seasonal component gives me the best result.

How can I use stlf or stlm functions to specify my order (3,1,6) in ARIMA method?

Currently, I am forecasting both components separately and merging them to obtain my forecasts.

Here is an example.

fit <- stlm(USAccDeaths, modelfunction=Arima, order=c(3,1,6)) fc <- forecast(fit) plot(fc) 

Similar Posts:

Rate this post

Leave a Comment