I have fit an ARMAX (an ARMA with exogenous/independent variable) model using Stata, and would now like to use said model to predict. My do-file looks a bit like this:
import delimited "data.csv"
generate n=_n
tsset n, generic
summarize
tsline sales
dfuller sales, regress trend
arima sales holiday avg_price, ar(1)
tsappend, add(50)
with sales being the depvar and holiday and avg_price being the indepvars.
After appending 50 new empty observations for predict, I would like to edit the observations I have for the indepvars, which are stored in a csv. Is this recommended, or even possible? Should a different approach be taken?
I have tried to use predict without editing the exogenous variables, which predictably gave me no results. I have also only found how to edit each data point manually with the data editor.