Show description

It is one of the great achievements of statistical decision theory that, although this principle is so general, it is nevertheless possible to give a precise prescription for fulfilling it. e. the one that minimizes the probability of an error of the second kind /? for a given probability for an error of the first kind a, is given by the Neyman-Pearson Lemma, after Jerzy S. Neyman (1894-1981) and Egon S. Pearson (1895-1980); see Rohatgi (1976). ,zn) of observations of the random variables (Zi,...

Gladitz and Willuhn (1982)). We don't in fact advocate such an approach, at least in the context of international agreements, as it implies some definite preconceived notion on the part of an inspectorate as to the a priori trustworthiness of the inspected party. 3 Several Classes 25 two players can be determined. We will assume that both protagonists behave non-cooper-atively, maximizing their expected payoffs independently, but at the same time taking their opponent's objective into account. In Chapter 9 it will be shown that it is reasonable for the inspector to announce his strategy to the inspectee in advance, and that it then suffices to determine that sample size distribution n which minimizes the non-detection probability under the assumption that the inspectee chooses his falsification strategy r so as to maximize the non-detection probability.

G. 3 should be rounded up to the nearest integer. If, however, the inspection procedure is repeated, they can be interpreted as expectation values to be achieved on the average. 3. 2 with e = 40 person-days. 73 An even simpler situation, to which we will be referring in the sixth chapter, arises if all the class-specific efforts and falsifications are equal, for example 01 = 02 = 80 tanks and 61 = €2 = 1 person-day. 34) and r* = r\ + r\ = 0/01, given by *i = jjNl9 n*2 = ^N2. 36) Thus the sample sizes are simply proportional to the total number of respective sites JVi and N2 and independent of the total number of undeclared tanks \i.