作者: M. J. Bayarri , M. H. Degroot
DOI: 10.1080/01621459.1991.10475135
关键词: Mathematics 、 Bayesian statistics 、 Bayesian average 、 Bayesian experimental design 、 Applied mathematics 、 Posterior probability 、 Bayesian probability 、 Statistics 、 Bayesian linear regression 、 Bayesian inference 、 Bayesian hierarchical modeling
摘要: Abstract Our goal is to study general properties of one Bayesian's subjective beliefs about the behavior another beliefs. We consider two Bayesians, A and B, who have different distributions for a parameter θ, Bayesian A's expectation B's posterior distribution θ given some data Y. show that when can take only values, always expects lie between prior B. Conditions are under which similar result holds an arbitrary real-valued θ. For vector we present useful expressions mean covariance matrix distribution. Examples illustrating relevance conditions results derived.