Title: A Bayesian framework for joint estimation of multiple networks
Authors: George Michailidis - University of Florida (United States) [presenting]
Abstract: A novel Bayesian approach is developed for joint estimation of multiple graphical models from a dictionary of possible/suggested ones. This problem arises in many applications, such as understanding co-expression networks from high-dimensional omics data obtained from different biological conditions or disease subtypes. We pursue a pseudo-likelihood based approach which provides robustness and computational efficiency. We establish strong posterior consistency and illustrate the efficacy of the proposed approach on both synthetic and real data.