How to Forecast the New Aggregate Live Naval Data?

Hiracan learning applies down two selar wals. Firstis regration, second is clasification. The hiracass is just sutuation, a describe. When i first t approach the retail problems, we look at the data with fagrat data. It tells you that sk u is tird, or stock keeping unique. Very specific description for item is sed in one second, and ant specific store. But if you do canditata on high aggregate label, you can forecast o reason - but it's very coner.