Abstract

INTRODUCTION The usual problem in Bayesian statistics is to try to guess the value of some vector quantity of interest g 2 R m , upon observing data X = x 2 R d under a parametric model X ¸ f(xj`), ` 2 \Theta ae R p , under the belief that ` ¸ ß(d`) and that g is related to the parameter by g = g(`) for some known function g : \Theta ! R m . Formally the problem is solved by Bayes' Theorem: ¯ g j E[g(`)jX = x] = R \Theta g(`)<F30.

Citations