A0916
Title: Comparing dependent directed Gaussian networks
Authors: Hongmei Zhang - University of Memphis (United States) [presenting]
Abstract: A Bayesian method is proposed to compare two dependent Gaussian Bayesian (directed) networks with topological ordering unknown. The method unifies topological ordering estimation, network construction, and network comparison. Simulations with different scenarios are used to assess the method. The approach is applied to epigenetic data measured at two-time points to test network differentiation and reconstruct the networks. Theoretical properties, simulation findings, and real applications support the effectiveness of the proposed approach.