CMStatistics 2018: Start Registration
View Submission - CMStatistics
Title: Design efficient composite likelihoods Authors:  Cristiano Varin - Ca Foscari University of Venice (Italy) [presenting]
Abstract: Composite likelihood is an inference function constructed by compounding component likelihoods based on low dimensional marginal or conditional distributions. Since the components are multiplied as if they were independent, the composite likelihood inherits the properties of likelihood inference from a misspecified model. The virtue of composite likelihood inference is combining the advantages of likelihood with computational feasibility. Given the wide applicability, composite likelihoods are attracting interest as scalable surrogate for intractable likelihoods. Despite the promise, application of composite likelihood is still limited by theoretical and computational issues that have received only partial or initial responses. Open theoretical questions concern characterization of general model conditions assuring validity of composite likelihood inference, optimal selection of component likelihoods and precise evaluation of estimation uncertainty. Computational issues concern how to design composite likelihoods to balance statistical efficiency and computational efficiency. After a critical review of composite likelihood theory, we shall focus on the potential merits of composite likelihood inference in modeling complex forms of dependence in discrete and categorical data.