A1223
Title: Bayesian high-dimensional mediation analysis incorporating neighborhood information
Authors: Yunju Im - University of Nebraska Medical Center (United States) [presenting]
Abstract: Mediation analysis is a useful tool to investigate the roles of the mediators that lie in the pathway from exposure to an outcome variable. With recent technological advances, researchers may encounter situations where the number of candidate mediators is high, necessitating mediator selection. In such cases, mediators may be correlated or exhibit network structures, making using ancillary information important in the mediator selection process. To address these challenges in a high-dimensional setting, a flexible statistical method is introduced that leverages external knowledge of the network structures between the mediators to improve selection and estimation accuracy. The proposed method's benefits are demonstrated through simulations and real-world data.
