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.
The sum of exponential-family statistics is sufficient for an IID sample
Statement
Assume AC. Let be standard Borel, let be a sigma-finite measure, and let be a nonempty exponential family of Exponential families and full rank with density . For each finite , the IID sample law has density relative to , and is sufficient for the sample. No full-rank, completeness, or converse characterization is asserted.
Facts & Assumptions
Given: AC, the standard-Borel sample space, sigma-finite , and a nonempty family with the displayed exponential-family density.
The Axiom of Choice is assumed by the common-reference factorization criterion and its conditional-law suppliers. The Axiom of Choice.
An exponential-family density is and integrates to one against . Exponential families and full rank.
Finite product sigma-algebras are generated by measurable rectangles, and finite products of sigma-finite measures are sigma-finite and satisfy the rectangle formula. The product sigma-algebra and its finite iterates For sigma-finite factors, the product measure exists, has the rectangle formula, is sigma-finite, and is unique.
Tonelli integrates nonnegative product densities by iterated integrals. Tonelli's theorem for nonnegative measurable functions on a sigma-finite product.
A nonnegative measurable density induces a measure by integration. The indefinite integral of a nonnegative measurable function is a measure.
Measures agreeing on a generating pi-system with a finite exhaustion agree on the product sigma-algebra. Measures agreeing on a generating pi-system are equal under an increasing finite-measure exhaustion from that pi-system.
The finite sample product of standard-Borel spaces is standard Borel, including the empty-product singleton; for , the Euclidean statistic space with its Borel sigma-algebra is standard Borel. Finite products of standard Borel spaces are standard Borel Finite-dimensional Euclidean space with its Borel sigma-algebra is standard Borel.
If every likelihood ratio relative to one in-model reference law is a Borel function of a statistic, the statistic is sufficient on standard-Borel spaces. Common-reference factorization criterion for sufficiency.
Sufficiency means there is one conditional-law kernel of the data given the statistic that works for every parameter. Sufficient statistics.
Finite-valued measurable functions are closed under finite sums and products. The closure proposition supplies sums, and the arithmetic theorem supplies products. Closure properties of measurable functions used by the integral Arithmetic and lattice operations preserve measurability whenever they are defined.
Proof
Given: The hypotheses in the Statement.
For fixed , each coordinate pullback of is measurable on the product sigma-algebra [F2], and the product density is measurable by [F9]. Its integral against is one by [F2], repeated Tonelli [F3], and . Thus is a probability measure by [F4]. On every measurable rectangle it equals by the rectangle formula and Tonelli, so the sigma-finite pi-system uniqueness theorem [F5] identifies it with the IID product law. Expanding gives the density displayed in the Statement.
Since is nonempty, choose . The density ratio on the common support is If any , both product densities vanish, so the same positive Borel function of satisfies everywhere. Integrating this identity on each measurable set proves that it is a version of the Radon–Nikodym derivative under .
By [F6] the finite sample space is standard Borel. If , [F6] also gives that the statistic target is standard Borel. If , is the empty product, a singleton with its full two-element sigma-algebra, which is standard Borel by the empty-product case of [F6]. In this case is the unique constant map, the dot products in step 2.1 are empty sums equal to zero, and its likelihood-ratio expression is still a Borel function of this singleton statistic. Thus both target cases meet the standard-Borel hypothesis. The reference belongs to the family, and step 2.1 makes every likelihood ratio a Borel function of . Applying [F7] gives a common conditional-law kernel, which is sufficiency by [F8]. The sole choice-axiom use is inherited from [A1] and [F7].
Depends on
- Exponential families and full rank
- Sufficient statistics
- The Axiom of Choice
- The product sigma-algebra and its finite iterates
- For sigma-finite factors, the product measure exists, has the rectangle formula, is sigma-finite, and is unique
- Tonelli's theorem for nonnegative measurable functions on a sigma-finite product
- The indefinite integral of a nonnegative measurable function is a measure
- Measures agreeing on a generating pi-system are equal under an increasing finite-measure exhaustion from that pi-system
- Finite products of standard Borel spaces are standard Borel
- Closure properties of measurable functions used by the integral
- Arithmetic and lattice operations preserve measurability whenever they are defined
- Common-reference factorization criterion for sufficiency
- Finite-dimensional Euclidean space with its Borel sigma-algebra is standard Borel
Used by
Nothing in the library uses this result yet.
Dependency tree · two levels
84 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
- Gelman et al., Bayesian Data Analysis, 3rd ed., §2.4 (standard reference, not scraped)
- Peter Kempthorne, MIT 18.655 Lecture Note 6: Sufficiency (standard reference, not scraped)