Dynamic pricing becomes a common practice nowadays in e-commerce such as hospitality and tourism industry, air-transportation or ride-share service. However, dynamic pricing algorithm is known to face the so-called cold start issue that no valid inferences can be drawn for users or items before sufficient information is fetched. In addition, pricing for product with high-dimensional features typically requires a fine tuning scheme on regularization to ensure good sale performance, enhance interpretation or even manage risk. In this talk, we study high dimensional dynamic pricing algorithm based on online Lasso procedure, where the customer transaction behavior is described by a structured choice model. We device a theoretical tool, termed time-uniform risk envelope, to manage the risk over the whole time horizon including the cold start period. This risk envelope result suggests an online regularization scheme that is adapted to different specific tasks in market demand dynamic such as maximizing the revenue. Regret of dynamic pricing algorithm with the online regularization scheme is developed and supported by simulations.
Bio: Guang Cheng is a Professor of Statistics at Purdue University. He received his PhD in Statistics from University of Wisconsin-Madison in 2006. His research interests include Big Data and Deep Learning. Cheng is the recipient of the NSF CAREER award, Noether Young Scholar Award and Simons Fellowship in Mathematics. He is currently a member of Institute for Advanced Study, Princeton in the Fall of 2019.