A0746
Title: Optimization-based framework for nonparametric empirical Bayes inference
Authors: Zhigen Zhao - Temple University (United States) [presenting]
Abstract: The purpose is to introduce a novel nonparametric empirical Bayes estimation method that approximates the unknown prior and, subsequently, the posterior using B-splines. The approach optimizes the prior by minimizing the maximum mean discrepancy between the empirical distribution and the model distribution. It provides a framework for nonparametric empirical Bayes inference using optimization. Extensive simulations and comparisons with traditional empirical Bayes methods demonstrate that the method delivers superior performance.
