A sequential reduction method for inference in generalized linear mixed models

• DOI: 10.48550/arxiv.1312.1903
• Publication Date: 2013-01-01
• Authors: Helen Ogden
• Keywords: Mixed model

The likelihood for the parameters of a generalized linear mixed model involves an integral which may be very high dimension. Because this intractability, many approximations to have been proposed, but all can fail when is sparse, in that there only small amount information available on each random effect. sequential reduction method described paper exploits dependence structure posterior distribution effects reduce substantially cost finding accurate approximation models with sparse structure.