This three-year award provides support for US-France cooperative research in statistical modeling. The collaboration involves the research groups of Martin Wells at Cornell University and Dominique Fourdrinier at the University of Rouen in France. The proposed research will examine the model selection problem in statistical modeling. The investigators will use several methodologies, in particular, a mixture of frequentist, Bayesian, and asymptotic methods. The ultimate goal is to determine estimators that have desirable properties for a large class of distributions. The US investigator brings to this collaboration expertise in Bayesian statistics and decision theory. This is complemented by the French investigator's expertise in classical mathematics and analysis. The project will advance understanding of modern statistical theory and practice.
