I'm trying to replicate a specification that uses a 3x9 seasonal filter. However, when I use udg() in the following object:
sa <- seas(x = ts(pib_headline_nsa_ts, start = c(2005, 1), end = c(2024,2), frequency = 4),
arima.model="(0 1 0)(0 1 1)", x11 = "", transform.function = "log", regression.aictest = c("easter"), outlier.types = 'NULL', forecast = c(maxlead = 6, maxback = 0))
The algorithm uses a 3x5 specification. Which parameter should I change to specify a 3x9 filter instead of automatic selection? I suppose it is in the x11 parameter, but I couldnt find information in any manual.
You can try specifying x11.seasonalma = "S3X5"
in the call:
library(seasonal)
sa <- seas(x = ts(pib_headline_nsa_ts,
start = c(2005, 1),
end = c(2024,2),
frequency = 4),
arima.model="(0 1 0)(0 1 1)",
x11.seasonalma = "S3X5", ## <----------- HERE
transform.function = "log",
regression.aictest = c("easter"),
outlier.types = 'NULL',
forecast = c(maxlead = 6, maxback = 0))
For example, using the AirPassengers dataset:
seas(AirPassengers, x11.seasonalma = "S3X5", arima.model = "(0 1 1)")
Call:
seas(x = AirPassengers, x11.seasonalma = "S3X5", arima.model = "(0 1 1)")
Coefficients:
Weekday MA-Nonseasonal-01
-0.003206 -0.325804