Alphabeta Math
TheoremStatement: Literature-sourcedProof: AI-adaptedPipeline-generated
How statement and proof provenance work

The first chip identifies the source of the statement or construction; the second identifies the source of its local proof or verification.

  • Literature-sourced: the exact statement appears in a cited source; only wording and notation differ.
  • AI-adapted: a semantically identical restatement of literature-sourced material, modulo indexing, notation, and boundary cases adopted by the library.
  • AI-generated: a genuinely novel statement formulated by AI, with no source for the claim itself.

These labels describe origin, not correctness: citations and verification chips remain separate evidence.

A full exponential family is complete

Statement

Assume AC and let (Pη)η∈Ω be a full-rank exponential family as in Exponential families and full rank, with Ω containing a nonempty open subset of Rk. Then its canonical statistic T is complete.

Facts & Assumptions

Given: the carrier, statistic, base measure, an open parameter subset, and a measurable real g such that Eη∣g(T)∣<∞ and Eηg(T)=0 for every η∈Ω.

[A1]

The Axiom of Choice: the transform-uniqueness lemma is used under its stated axiom branch.

[F1]

Exponential families and full rank: the law has density heη⋅T−A(η), with finite positive normalizer on Ω.

[F2]

Complete statistics: to prove completeness it suffices to show the stated mean-zero function vanishes almost surely for every parameter.

[F3]

Uniqueness of a multivariate exponential transform: a finite signed measure with an exponential moment and zero Laplace transform on a real neighborhood is zero.

[F4]

A real L^1 density defines a finite signed measure with its canonical Hahn and Jordan data and The total variation |nu|(E) from countable measurable partitions: an integrable real density defines a finite signed measure with variation density ∣g∣, and variation is characterized by countable partition sums.

[F5]

Every nonnegative measurable function admits an explicit increasing sequence of simple approximations and Monotone convergence for the integral: nonnegative measurable functions have increasing simple approximations whose integrals converge to their integral.

Proof

1.1F1F4givenconstruct

Choose η0 in the open subset and ϵ>0 so the closed cube η0+[−ϵ,ϵ]k lies in Ω. Define the finite signed measure ρ(A)=∫Ag(T(x)) Pη0(dx) by [F4], and set ν(B)=ρ(T−1B) for Borel B⊆Rk. Since T is measurable, preimages preserve disjoint countable unions, so ν is a finite signed Borel measure. The preimages of a countable partition form a countable partition, giving ∣ν∣≤T#∣ρ∣ by [F4]. Hence ∣ν∣(Rk)≤Eη0∣g(T)∣<∞.

2.1F1F3F4F5step 1.1givenalgebra

Choose 0<δ<ϵ. Since eδ∥s∥1≤∑σ∈{−1,1}keδσ⋅s and ∣ν∣≤T#∣ρ∣ by step 1.1, the weighted-integral bound follows first for nonnegative simple functions from that setwise domination and then for eδ∥s∥1 by [F5]. For each sign vector, changing from Pη0 to Pη0+δσ contributes the factor Z(η0+δσ)/Z(η0), so ∫eδ∥s∥1∣ν∣(ds)≤∑σZ(η0+δσ)Z(η0)Eη0+δσ∣g(T)∣<∞, the exponential moment required by [F3].

3.1A1F3step 2.1givenalgebra

For every real t with ∥t∥∞<δ, its Laplace transform is ∫et⋅sν(ds)=Z(η0+t)Z(η0)Eη0+tg(T)=0 by the mean-zero premise. The transform-uniqueness result [F3], under axiom branch [A1], gives ν=0.

4.1F1F2step 3.1givenalgebra∎

For every Borel B, 0=ν(B)=ρ(T−1B)=∫Bg(s)Pη0T(ds) by the definition of the pushforward measure and ρ. Taking B={g>0} and B={g<0} shows both positive and negative parts have zero integral, so g(T)=0 Pη0-almost surely. Every family density is strictly positive on {h>0}, so all Pη are equivalent to Pη0 and the same vanishing holds for every η. By [F2], T is complete.

Depends on

Used by

Dependency tree · two levels

42 results within two dependency steps of this one, each drawn at its shortest distance from it. An arrow runs from a result to what uses it, so the chart reads left to right and ends at this result, which carries a heavier outline. Every node is a link to that result. Click elsewhere on the chart to enlarge it.

Sources