报告题目 | The Boundaries of Personalization in Repeated Customer Interactions |
报告人（单位） | 高品（香港中文大学（深圳）） |
主持人（单位） | 李四杰、陈静（东南大学） |
时间地点 | 2023年4月28日（周五）上午10：00 腾讯会议：766-141-035 |
报告摘要和内容： This paper presents the Repeated Purchase Multinomial Logit model, which analyzes the decision-making process of a representative customer across several stages. In each stage, the customer selects a product from the available options or opts for the outside option that provides the highest utility. The seller's recommendation strategy can be personalized or non-personalized, depending on whether it is contingent on the customer's purchase history. The study aims to investigate the performance ratio between the optimal non-personalized and personalized recommendation strategies. We start by considering the two-period problem. It demonstrates that when the available sets of products in both periods are identical, the optimal personalized recommendation, exhibiting some elegant structures, can be identified with ease. In contrast, when the available sets of products in both periods are disjoint, finding the optimal personalized recommendation becomes NP-hard, and a fully polynomial time approximations scheme is devised. The non-personalized performance ratio in the two cases is identified to be 2/3 and 4/5, respectively. For the general multiple-period problem where product utilities may also evolve over time, the non-personalized performance ratio remains above 1/2. Furthermore, if the available sets of products in any two periods are disjoint, the ratio is enhanced. | |
报告人简介：
Gao Pin is currently an assistant professor at the School of Data Science, The Chinese University of Hong Kong, Shenzhen. He is also affiliated with Shenzhen Institute of Artificial Intelligence and Robotics for Society. He received a B.S. in physics from Wuhan University in 2013 and a M.Phil. in physics from Hong Kong University of Science and Technology (HKUST) in 2015, after which he worked in industry for two years. In 2021, he received a Ph.D. in Industry Engineering and Decision Analytics from HKUST. Dr. Gao's research interests include revenue management and operations management. | |
