Bayes' Approach to System Random Inspections for Nuclear Materials Control and Accounting

A general approach, based on Bayes’ theorem, to define a random sample size that would assure meeting a preset statistical criterion is presented here. The approach uses the hypergeometric probability distribution and facilitates the introduction of a function that describes the possible defect distribution in a system (the socalled binary distribution function or BDF) and accounting for a priori information, if any. The method will allow correct use of the statistical information about the system that was accumulated as a result of previous sampling measurements. In the particular case of absence of the a priori information about the system, it is shown that both the aforementioned methodology and a typical statistical hypothesis method lead to similar results. The method is applicable to many statistical tasks in nuclear material control and accounting, in particular during inspections.