Exchangeable Belief Structures
本文提出用内积空间（信念结构）组织不确定性陈述，反对贝叶斯学派用可交换性证明频率统计模型，转而通过可交换与共可交换信念结构推导表示定理及统计模型。
Abstract Abstract The static features of beliefs, namely those structures that relate to one's beliefs at a particular moment, are described. It is argued that expectation (or prevision) should be the fundamental quantification of individual statements of uncertainty and that inner product spaces (or belief structures) should be the fundamental organizing structure for collections of such statements. Objections are given to the usual Bayesian justification of relative frequency-based statistical models via exchangeability. I offer an alternative approach, via exchangeable and co-exchangeable belief structures, and derive the representation theorem, and thus the implied statistical models, for these structures. Key Words: Linear methodsPrevisionRelative frequencyRepresentation theorems