Answers are grounded in this paper and cite their sources.
The following lemma proved by Neyman and Pearson [1] is basic in the theory of testing statistical hypotheses: LEMMA. Let f₁(x), ⋯, fₘ₊₁(x) be $m + 1$ Borel measurable functions defined over a finite dimensional Euclidean space R such that ∫R |fᵢ(x)|dx < ∞ (i = 1, ⋯, m + 1). Let, furthermore, c₁, ⋯, cₘ be m given constants and S the class of all Borel measurable subsets S of R for which (1.1) ∫S fᵢ(x) dx = cᵢ \\ (i = 1, ⋯, m). Let, finally, S₀ be the subclass of S consisting of all members S₀ of S for which (1.2) ∫S₀ fm + 1(x) dx ∫S fₘ₊₁(x) dx for all S in S. If S is a member of S and if there exist m constants k₁, ⋯, kₘ such that (1.3) fm + 1(x) k₁f₁(x) + ⋯ + kₘfₘ(x) when x ε S, (1.4) fm + 1(x) k₁f₁(x) + ⋯ + kₘfₘ(x) when x ε S, then S is a member of S₀. The above lemma gives merely a sufficient condition for a member S of S to be also a member of S₀. Two important questions were left open by Neyman and Pearson: (1) the question of existence, that is, the question whether S₀ is non-empty whenever S is non-empty; (2) the question of necessity of their sufficient condition (apart from the obvious weakening that (1.3) and (1.4) may be violated on a set of measure zero). The purpose of the present note is to answer the above two questions. It will be shown in Section 2 that S₀ is not empty whenever S is not empty. In Section 3, a necessary and sufficient condition is given for a member of S to be also a member of S₀. This necessary and sufficient condition coincides with the Neyman-Pearson sufficient condition under a mild restriction.
No takes yet. Share an insight, caveat, or question.
Dantzig et al. (1951) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: