It seems the current TBATS implementation does not allow external regressors. Still, I think (maybe I am wrong 🙈 ) that it should be "easy" as we need to pass X here https://github.com/Nixtla/statsforecast/blob/main/statsforecast/tbats.py#L875 into the auto_arima_f as this function already have and xreg argument. Do you think it is that simple? If so, I could open a PR.
Use case
Having the availability of including external regressors can be very useful for real applications.
Description
It seems the current TBATS implementation does not allow external regressors. Still, I think (maybe I am wrong 🙈 ) that it should be "easy" as we need to pass
X
here https://github.com/Nixtla/statsforecast/blob/main/statsforecast/tbats.py#L875 into theauto_arima_f
as this function already have andxreg
argument. Do you think it is that simple? If so, I could open a PR.Use case
Having the availability of including external regressors can be very useful for real applications.