Dr. Wang Zuocheng's Lecture

Post Time:2016-12-23

Time: Jan. 9th, 2017     10:00   a.m.

Room:   Information Building 0224

Lecturer:Zuozheng Wang received his Ph.D. in supply chain management from Robert H. Smith School of Business at University of Maryland. Previously, he held a master’s degree from Singapore-MIT Alliance, a B.E. and M.E. from Shanghai Jiaotong University. Dr. Wang is currently working as a consultant of data analytics at a multinational development bank. His main research interest is in empirical operations management and business analytics. Recently, Dr. Wang focuses on investigating how consumer choices impact operations management decisions.


Title: Estimating Choice Models with Censored Sales Data



Discrete choice models have been used to understand demand for products and services in operations and revenue management models. In many cases, a firm only has its own sales data, while it is difficult to get the sales data of competing products. Thus, it calls for a robust estimation method that works well even with limited data. Existing estimation methods are all based on some special assumptions and may result in unreliable and biased estimates. We develop an approach for estimating customer preferences based on the case-control sampling with incomplete data. This problem is estimated by using the expectation-maximization (EM) algorithm. Our method is shown to be superior to all existing methods using real data from a hotel chain. Simulations demonstrate that our method can reduce estimation errors from 16% to 1% and provide valid inputs for revenue management systems.