报告题目：Significance Test for Fine-Grained Dynamic Pricing: High-dimensional Consumer Features and Nonparametric Price Sensitivity
报告时间： 2025年05月21日（星期三）16:00-16:40
报告地点：翡翠湖校区科教楼B座1005会议室
报 告 人：董枘朋（中国科学技术大学）
主办单位：合肥工业大学经济学院
报告简介：
We analyze the pricing problem faced by a retailer selling a product to a sequence of customers. The retailer can dynamically charge a different personalized price to each potential customer by leveraging the large amount of information it has about the individual consumer and the purchases by previous consumers while considering the influence of overall market price sensitivity. We assume that (i) the number of consumer features is large in comparison with previous demand observations and (ii) the overall market price sensitivity can be linear or nonlinear-and unknown to the retailer. The retailer aims to maximize the cumulative revenue within a finite time horizon, which requires balancing between demand learning and revenue maximization. We propose a dynamic pricing algorithm based on a two-stage semi-parametric demand learning framework: in the first stage we account for the high-dimensionality of consumer features and their interplay with price and demand; in the second stage we perform a non-parametric estimation of the overall market price sensitivity. We do so by minimizing the regret of the overall proposed policy. For the first time in the personalized dynamic pricing literature, we are able to provide a formal statistical test for the learned demand parameters. Our method not only provides a significance test for the model but also endows the managers with a theoretical confidence for their fine-grained personalized pricing in practice. Finally, we perform extensive numerical experiments to illustrate that our method does outperform the existing methods in terms of regret performance.
报告人简介：
董枘朋，中国科学技术大学管理学院博士后研究员。本科毕业于西南交通大学，博士毕业于中国科学技术大学，主要从事高维统计推断、变量选择、低秩矩阵估计等统计方法研究，并探索统计方法在商科领域的交叉应用。其研究成果发表于INFORMS Journal on Computing、Journal of Machine Learning Research、Computational Statistics & Data Analysis、Statistics and Computing、Statistics & Probability Letters 等国际权威期刊。目前主持国家自然科学基金青年项目（C类）和中国博士后科学基金面上项目，并荣获中国科学技术大学墨子资助。
