-
Notifications
You must be signed in to change notification settings - Fork 11
Description
Hi everyone
I got the time series Problem.
I want to use multivar. GARCH model with binary predictor, where binary (dummy) variable being calendar feature like January (1 for Yes 0 for no)
Problem: I cannot fit calendar features as dummy variable.
Could someone suggest any useful packages with code example.
I tried to do it with garchmodells package.
Below is my code and error message:
DAX_rugarch_01_month <-
garch_multivariate_reg(
mode = "regression", type = "ugarchspec"
) %>%
set_engine("rugarch", specs= list(
spec1 = list(mean.model = list(armaOrder = c(1,0))),
spec2 = list(mean.model = list(armaOrder = c(1,0))),
spec3 = list(mean.model = list(armaOrder = c(1,0)))
)) %>%
fit(Return~ Date + January, data = DAX_train_January)
Error:
multifit GARCH-->error: speclist length not equal to data length
Somethiny is off with my specs at set_engines? The number of dimensions is 3, so I got spec1, spec2, spec3.
Would appreciate any help!