Search |  Contact |  SRI Home Do not follow this link, or your host will be blocked from this site. This is a spider trap. Do not follow this link, or your host will be blocked from this site. This is a spider trap. Do not follow this link, or your host will be blocked from this site. This is a spider trap.A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A ASRI International.  333 Ravenswood Avenue.  Menlo Park, CA 94025-3493. SRI International is a nonprofit corporation.

AIC Seminar Series

Bayesian Model Specification

David DraperUniversity of California, Santa Cruz[Home Page]

Notice:  hosted by Sugato Basu

Date:  Wednesday, June 14th 2006 at 10:30am

Location:  EJ228  (Directions)


A standard (data-analytic) approach to statistical model specification, practiced with equal vigor by Bayesians and non-Bayesians alike, involves the initial choice, for the structure of the model, of one or another of a variety of standard parametric families, followed by modification of this initial choice – once data begin to arrive – if the data suggest deficiencies in the original specification.

In this talk (a) I will argue that this approach is formally incoherent, because it amounts to using the data both to specify the prior distribution on structure space and to update using this data-determined prior; (b) I will identify two approaches to avoiding (at least in principle, and with a fair amount of data) the incoherence in (a) – (1) Bayesian nonparametric modeling and (2) three-way out-of-sample predictive validation; (c) I will provide details on implementing (2); (d) I will argue that to make progress in coherent Bayesian model specification in complicated problems you have to either implicitly or explicitly choose a utility structure which defines, for you, when the model currently being examined is good enough; (e) I will argue that it is best to make this choice explicitly on the basis of real-world considerations regarding the use to which the model will be put; and (f) I will contrast model selection methods based on the log score and deviance information criteria (DIC) as two examples of (e) with utilities governed by predictive accuracy. This is joint work with a former Ph.D. student, Milovan Krnjajic.

Keywords: Bayesian model specification, DIC, model selection as a decision problem, predictive log scoring rule, three-way out-of-sample predictive validation, Bayesian nonparametric modeling

   Bio for David Draper

David Draper is a Professor in, and Chair of, the Department of Applied Mathematics and Statistics in the Baskin School of Engineering at the University of California, Santa Cruz. From 2001 to 2003 he served as the President-Elect, President, and Past President of the International Society for Bayesian Analysis (ISBA). His research is in the areas of Bayesian inference and prediction, model uncertainty and empirical model-building, hierarchical modeling, Markov Chain Monte Carlo methods, and Bayesian nonparametric methods, with applications mainly in medicine, health policy, education, and environmental risk assessment. He has a particular interest in the exposition of complex statistical methods and ideas in the context of real-world applications.

   Note for Visitors to SRI

Please arrive at least 10 minutes early as you will need to sign in by following instructions by the lobby phone at Building E (or call Wilma Lenz at 650 859 4904, or Eunice Tseng at 650 859 2799). SRI is located at 333 Ravenswood Avenue in Menlo Park. Visitors may park in the parking lots off Fourth Street. Detailed directions to SRI, as well as maps, are available from the Visiting AIC web page. There are two entrances to SRI International located on Ravenswood Ave. Please check the Building E entrance signage.

SRI International
©2018 SRI International 333 Ravenswood Avenue, Menlo Park, CA 94025-3493
SRI International is an independent, nonprofit corporation. Privacy policy