A0774
Title: Shapley value and DOE
Authors: Wei Zheng - University of Tennessee (United States) [presenting]
Abstract: The Shapley value is a well-known concept in cooperative game theory that provides a fair way to distribute revenues or costs among players. It has found applications in many fields besides economics, such as marketing and biology. Recently, it has been widely applied in data science for data quality evaluation and model interpretation. 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 evaluating the values for $2^n$ different coalitions, which makes it infeasible for large $n$. A typical approach to address this issue is to take a sample of the subsets or permutations of players to approximate the Shapley values. However, with the inspiration of principles in the design of experiments, it is argued that a well-designed sampling mechanism would outperform random sampling schemes. Two types of design structures are introduced that can help with the computation, as well as the inference problem through the lens of Bayesian factorial designs, which is difficult in view of scarce observations from the design point of view.
