SARIMA Model multi-step forecast evaluation issue

[ketyi]

Hi @ritvikmath,
First, your work provided for the community is far more than excellent & super valuable.

In SARIMA Model.ipynb the workaround to improve on the residuals might be problematic:
the original task is a 6-item suffix forecast but the relaxation of that to the average of 1-step forecasts is not really focusing on the same task IMHO.