Walmart Inc.
Predictive shopping

Abstract:

A budget-constrained, machine-learning system is described that creates a shopping (purchase) list and performs on-line ordering and delivery. It receives the shopper's past purchase receipts from a retail store, pharmacy and/or auto center. It may attach to a web server to acquire on-line browsing information. The system creates a Purchase List from acquired information. The system receives a budget and determines if all items on the Purchase List can be bought under the budget. If not, the items are given priority ratings. The system walks down the list to in decreasing priority rating order identifying items to purchase without exceeding the budget. The shopper may override the items identified to be purchased. Shopper override is monitored by a machine learning engine which adjusts the priority rating of the item or the period of replacement for the next shopping trip/session, allowing for more accurate results and flexibility.