How can we calculate forecast at aggregated level and desaggregate after?

Content

Due to the fact that there is a high number of items on each category and each item has a shipment history spread through the history time buckets, customer wants to load history at item level, but calculate the forecast at an aggregated level and only after determining the forecast at this aggregated level (category in this case), the forecast should be decomposed at item level.

 

In customer experience, this type of calculation normally brings a higher accuracy and a lower mape as the trends on the forecast are being calculated at an aggregated level that tends to be more regular and easier to determine statistically.