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.
Hellinger invariance and comparison with library total variation
Statement
Assume AC for the imported density-representative convention. Let be probability laws on a measurable space, both dominated by one specified common sigma-finite positive measure. Then the root-L2 Hellinger distance from Root-L2 Hellinger distance for dominated laws is independent of the choice of common dominating measure. With the library convention from Total variation distance for probability laws,
Facts & Assumptions
Given: Probability laws with a specified common sigma-finite dominating measure. The conclusions are the common-measure invariance and both displayed inequalities.
Assume the Axiom of Choice for the general density-selection convention in Likelihood densities. After finite-valued density representatives are supplied, the calculations use no further choice. The Axiom of Choice
For finite-valued nonnegative density representatives against a common measure , . Root-L2 Hellinger distance for dominated laws
The library total-variation distance is , with no factor in the event supremum. Total variation distance for probability laws
A common dominating measure is sigma-finite and satisfies . Dominated statistical experiments
Under AC, a law dominated by the specified sigma-finite measure has a finite-valued nonnegative measurable density representative. Likelihood densities
For square-integrable real functions , . Holder's inequality for integrals, including the endpoint cases
A finite measure is sigma-finite, and a sigma-finite measure has a countable finite-measure exhaustion. Finite, sigma-finite, and semifinite measures
Sums, products, scalar multiples, and lattice operations of finite real-valued measurable functions are measurable. Arithmetic and lattice operations preserve measurability whenever they are defined
A set defined by a real threshold of a measurable real-valued function is measurable. Threshold characterisations of real-valued and extended-real-valued measurability
For nonnegative measurable , the density measure satisfies . Integrating against a density agrees with integrating the product
The nonnegative integral is additive, monotone and homogeneous. Additivity of the nonnegative Lebesgue integral, Monotonicity and nonnegative homogeneity of the nonnegative integral
Proof
Given: The hypotheses in the Statement.
By [A1], [F3], and [F4], take finite-valued nonnegative densities and for a common sigma-finite dominator . Define . For disjoint measurable , countable additivity of and regrouping nonnegative series give ; also and . Thus is a finite, hence sigma-finite, measure by [F6]. Additivity of the nonnegative integral [F10] gives , with measurable by [F7].
Using against , the set is measurable by [F7,F8]. On , , so additivity [F10] gives . On , and the total masses are one, hence . Additivity then gives . For any measurable , the difference on lies in by monotonicity [F10], while the difference on lies in ; thus , and attains . Therefore [F2] gives .
On define and , and set both to zero on . For , is the whole space; for , it is . Thus [F7,F8] make measurable, and the same argument applies to . Since where , [F9] gives and for every measurable . Hence are finite-valued density representatives of against .
Pointwise . Hölder [F5] therefore gives . The squared second factor is , since [F5] applied to bounds their product integral by . Step 1.2 now gives .
Write for the defining Hellinger integral computed from supplied densities against a common dominator . Applying [F1] and [F9] to the densities in step 2.1 gives , because both original densities vanish where . For any measure and finite-valued nonnegative densities of the same finite measure with respect to , they agree -almost everywhere: on , equality of their density integrals gives , so . Since , and the same argument with swapped handles , this proves uniqueness almost everywhere. Thus changing supplied representatives against a fixed dominator does not change the Hellinger integral.
For any other common sigma-finite dominator , repeat steps 1.1 and 2.1 using its densities to obtain density representatives against the same finite measure . The uniqueness argument in step 3.1 gives and -almost everywhere. The same density-change calculation as in step 3.1 therefore identifies the Hellinger integral against with the same canonical value against , and hence with that against .
Pointwise . Integrating by monotonicity [F10] and using [F1] and step 1.2 gives .
Depends on
- Root-L2 Hellinger distance for dominated laws
- Total variation distance for probability laws
- Dominated statistical experiments
- Likelihood densities
- Probability measures and probability spaces
- Holder's inequality for integrals, including the endpoint cases
- The Axiom of Choice
- Finite, sigma-finite, and semifinite measures
- Arithmetic and lattice operations preserve measurability whenever they are defined
- Threshold characterisations of real-valued and extended-real-valued measurability
- Integrating against a density agrees with integrating the product
- Additivity of the nonnegative Lebesgue integral
- Monotonicity and nonnegative homogeneity of the nonnegative integral
Used by
Dependency tree · two levels
48 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
- B. J. K. Kleijn, Lectures on Nonparametric Bayesian Statistics, §8.1 equation (8.2) and §9.4 Lemma 9.8 (standard reference, not scraped)