McMASTER UNIVERSITY STATISTICS SEMINAR

Week of November 1 - 5, 1999

SPEAKER:	Dr Michael Evans Department of Statistics, University of Toronto
TITLE:	"Bayesian Inference and the Concept of Surprise"
DAY:	Wednesday, November 3, 1999
TIME:	3:30 p.m. [Coffee & cookies in BSB-202 at 3:00 p.m.]
PLACE:	BSB-108

SUMMARY

The argument is put forward that the concept of surprise, and more particularly relative surprise, can serve as a logical basis for deriving inferences. A virtue of the relative surprise approach is that all inferences; e.g. estimates, credible regions, tests of hypotheses, prediction, model checking, etc. are derived via basically the same argument. Further all of these inferences are completely invariant under transformations and it is argued that this is a natural requirement for any acceptable theory of inference. Several controversies concerning statistical inference, and the relevance of the relative surprise approach to these, are discussed.

ABOUT THE SPEAKER

Mike Evans graduated with a B.Sc. from the University of Western Ontario in 1974 and with a Ph.D. from the University of Toronto in 1977. He has been employed there since graduating, with leaves spent at Stanford and Carnegie Mellon. He served as Chair of the Department of Statistics during 1992-97. He has been an Associate Editor of the Journal of the American Statistical Association since 1989 and an Associate Editor of The Canadian Journal of Statistics since 1997. He is interested in the foundations of statistical inference, Bayesian inference methodology and methods for approximating integrals. Together with Tim Swartz, he has a forthcoming book from Oxford University Press called Approximating Integrals via Monte Carlo and Deterministic Methods.

REFERENCES

The reference below, which Dr Evans has provided as useful background for his talk, has been placed on reserve at Thode Library (STATS 770: Statistics Seminar):

[1] Evans, M. (1997), "Bayesian Inference Procedures Derived Via the Concept of Relative Surprise," COMMUNICATIONS IN STATISTICS - THEORY AND METHODS, Vol. 26, No. 5, pp. 1125-1143.