统计缺省扩充的两种算法

Two Kinds of Algorithms of Statistical Extensions

讨论统计缺省理论的扩充,它是经典缺省逻辑(R e iter缺省)的推广,它借助错误参数ε,允许我们在标准的推理统计中模型化普遍的推理模式。由于统计缺省扩充不具备封闭性,在其对应的R e iter扩充存在时,它的扩充可能不存在;甚至可能出现正规的统计缺省理论没有扩充的情形。但是如果给出一定的限制条件,就可以克服这一缺陷。本文给出了在这种限制条件下,统计缺省理论扩充的两种计算方法。

This paper discusses statistical extensions of statistical default logic, an expansion of classical （i. e. Reiter＇ s） default logic that allows us to model common inference patterns found in standard inferential statistics, which is an expansion with an error-parameter ε. The present paper first points out that statistical default the ory does not possess closeness property, it may have no extensions while the corresponding Reiter＇s default theory does have an extension. Moreover, statistical default may have no ε-extensions even if it is a normal default. Under certain restricted conditions the present paper proposes two kinds of method for constructing extensions of statistical default theory.