报告人：Zheng Wei
报告地点：数学与统计学院415教室
报告时间：2024年06月25日星期二10:00-11:00
报告摘要：
Shapley value is a well-known concept in cooperative game theory which provides a fair way to distribute the revenues or costs to each player. Recently, it has been widely applied in various fields such as data science, marketing, and genetics. However, the computation of the Shapley value is an NP-hard problem. For a cooperative game with $n$ players, calculating Shapley values for all players requires calculating the value function for $2^n$ different coalitions, which makes it infeasible for a large $n$. In this paper, we propose a fast approximation approach for Shapley values based on fractional factorial designs. Under certain conditions, our approach can obtain true Shapley values by calculating values of fewer than $4n^2-4$ different coalitions. In general, highly accurate approximations of Shapley values can also be obtained by calculating values of additional $O(n^2)$ different coalitions. Multiple simulations and real case examples demonstrate that, with equivalent computational cost, our method provides significantly more accurate approximations compared with several popular methods.
主讲人简介：
Dr. Wei Zheng is an associate professor in Department of Business Analytics and Statistics at University of Tennessee. He obtained his PhD in Statistics at University of Illinois at Chicago. He is currently an associate editor for Statistica Sinica, Journal of Statistical Planning and Inference, and Metrika. His research interest is in design of experiment (DOE). He is also interested in the interaction between DOE and machine learning. His research has been published in Annals of Statistics, Journal of American Statistical Association, Statistica Sinica, etc.
