贝叶斯决策理论的理性观及其相对性

The Rationality and Its Relativity in Bayesian Decision Theory

贝叶斯决策理论的基础在于如何表达价值和信念。该理论基于一些偏好公理而得到效用函数和信念度函数(分别表达价值和信念)。这些函数的性质可以表述为两个表达定理,它们展现了贝叶斯决策理论的理性观:当事人的价值及信念度赋值必须都是一致的。然而,价值及信念度赋值的一致性不仅取决于偏好公理,而且还依赖于效用单元的选择。

The foundation of Bayesian decision theory is how to represent our values and beliefs.The utility and degree of belief functions,which represent values and beliefs respectively,can be constructed based on some preference axioms.And the qualities of these functions can be represented by two representation theorems,which lay out the principles of the rationality in Bayesian decision theory: the assignments of values and degrees of belief must be coherent.However,the coherency is depended not only on the preference axioms but also on the choice of which prizes count as constants in terms of utility.