Improving Sales Forecasting Accuracy: A Tensor Factorization Approach with Demand Awareness

• Activity Number: 392 - Recent Advances in Tensor Learning
• Type: Invited
• Date/Time: Thursday, August 12, 2021 : 2:00 PM to 3:50 PM
• Sponsor: IMS
• Author(s): Xuan Bi* and Gediminas Adomavicius and William Li and Annie Qu
• Companies: University of Minnesota and University of Minnesota and Shanghai Jiao Tong University and Unviersity of California Irvine
• Keywords: Tensor factorization; Sales competition; Long short term memory; End-to-end learning

Abstract

Due to accessible big data collections from products and stores, advanced sales forecasting capabilities have drawn great attention from many companies especially in the retail business because of its importance in decision making. Improvement of the forecasting accuracy, even by a small percentage, may have a substantial impact on companies' production and financial planning, marketing strategies, inventory controls, supply chain management, and eventually stock prices. Specifically, our research goal is to forecast the sales of products in stores. Motivated by tensor factorization methodologies for context-aware recommender systems, we propose an Advanced Temporal Latent-factor Approach to Sales forecasting, which achieves accurate and individualized prediction for sales. Our contribution is to incorporate sales competition under the tensor framework. We further extrapolate a tensor into future time periods using seasonal ARIMA and LSTM models. The advantages are demonstrated on eight product category datasets collected by the Information Resource, Inc., where a total of 165 million weekly sales transaction from more than 1,500 grocery stores over 15,560 products are analyzed.