Title: Approximate likelihoods

Speaker: Nancy Reid

Date: August 14
Time: 8:30 - 9:30
Room: Ballroom A

Chair: Maria J. Esteban

Abstract: In complex models likelihood functions may be difficult to compute, or depend on assumptions about high order dependencies that may be difficult to verify. A number of methods have been devised to compute inference functions either meant to approximate the true likelihood function, or to provide inferential summaries that balance statistical efficiency with ease of computation. Examples include variational approximations, composite likelihood, quasi-likelihood, indirect inference, and Laplace-type approximations. This talk will survey various approximations to likelihood and likelihood inference, with a view to identifying common themes and outstanding problems.


Key Features:
Awarding ceremony of ICIAM prizes
Invited lectures
Prize lectures
Industrial Minisymposia
Contributed Minisymposia
Poster sessions
Embedded and satellite meetings
Public outreach sessions