I've been working to switch some code over to statsforecast for baseline forecasting.
I've hit a wall with integrating past covariates (date/uniuqe_id dependent features that are only known up to the prediction time). Future covariates seem to work effortless - you simply cut them at prediction time and merge the past to training data and feed the future part as df_X.
Are past covariates simply not supported or am I just not seeing something obvious🙂?
Thanks for your help! 🙌
@fede (nixtla) (they/them) sorry for tagging you directly, but would you mind chiming in?
08/14/2023, 12:22 AM
Azul is on vacation, so I’m answering questions. If you want to use features that are only known until prediction time, I am wondering what you would use for forecasting when they are unknown?
Wondering if you have an idea of how they should be handled? Would you get the lag or something like that? Would love to hear your thoughts