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 distance is a metric and controls L1 and total variation
Statement
For probability densities relative to one fixed measure , define as in Root-L2 Hellinger distance between density laws and set . Then is a metric on the induced probability laws and where library total variation is the event supremum .
Facts & Assumptions
Given: A measure space and nonnegative measurable densities with integrals one; let be their induced probability laws.
Root-L2 Hellinger distance between density laws: for this fixed dominating measure.
Minkowski's inequality for integrals, including : for functions, .
Holder's inequality for integrals, including the endpoint cases: when both functions are in .
A nonnegative measurable function has integral exactly when it vanishes almost everywhere: a nonnegative measurable function has zero integral exactly when it vanishes -almost everywhere.
The Lebesgue integral is linear on : integration is linear for integrable functions; here because , both densities integrate to one, and [F9] plus [F10] bound the integral of by .
Arithmetic and lattice operations preserve measurability whenever they are defined: sums, differences, products, and absolute values of measurable functions are measurable.
Composition with a Borel measurable outer map preserves measurability: Borel functions of measurable real-valued functions are measurable; in particular this applies to square root on , the sign function, and threshold indicators.
Monotonicity and nonnegative homogeneity of the nonnegative integral: the nonnegative integral preserves pointwise order.
Additivity of the nonnegative Lebesgue integral: the nonnegative integral is additive on nonnegative measurable functions.
Proof
Since are measurable, [F8] makes measurable and [F7] makes their difference and absolute value measurable. Their unit integrals put both roots in . The triangle inequality in [F2] gives symmetry, nonnegativity, and the triangle inequality for . If , [F5] gives almost everywhere, hence almost everywhere and . Conversely, put by [F6]. If , then for every measurable , by [F6]. For each integer , the set is measurable by [F8]. Define ; it is nonnegative and measurable by [F7], and linearity gives . By [F5], almost everywhere; since on , . Applying the same argument to shows . Since is the countable union of these null sets, almost everywhere and . Thus is a metric on these laws.
Put and , which is measurable by [F7, F8]. Since , the positive and negative parts have equal integral, so . For every measurable , positivity of on and nonpositivity on , together with monotonicity [F9], give ; equality in the absolute supremum is attained at (or ). Thus [F4] gives .
The functions in the pointwise inequality are measurable by [F7, F8]. If , then , because when the denominator is positive; the case is symmetric, and when both vanish the inequality is zero on both sides. Monotonicity [F9] and [F1] give .
By [F7, F8], the displayed products and sums of square roots are measurable. Factor . By [F3], . Also , again by [F3] and the unit norms. Hence , and step 1.2 gives .
Depends on
- Root-L2 Hellinger distance between density laws
- Total variation distance for probability laws
- Minkowski's inequality for integrals, including $p = \infty$
- Holder's inequality for integrals, including the endpoint cases
- A nonnegative measurable function has integral $0$ exactly when it vanishes almost everywhere
- The Lebesgue integral is linear on $L^1(\mu)$
- Arithmetic and lattice operations preserve measurability whenever they are defined
- Composition with a Borel measurable outer map preserves measurability
- Monotonicity and nonnegative homogeneity of the nonnegative integral
- Additivity of the nonnegative Lebesgue integral
Used by
- Countably partitioned sieves yield posterior contraction by weighted entropy Lemma
- Hellinger entropy for compact-location normal mixtures in a scale band Lemma
- Sieve metric entropy produces uniform exponential tests Lemma
- The dyadic scale bands satisfy the weighted entropy condition Lemma
- The mixture prior places local KL and log-square mass near a compact smooth truth Lemma
- IID posterior contraction by entropy and prior mass Theorem
Dependency tree · two levels
39 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
- Ghosal, Ghosh and van der Vaart, Convergence rates of posterior distributions, §2 (standard reference, not scraped)
- Ghosal, The Dirichlet Process, Related Priors and Posterior Asymptotics, §2.5 (standard reference, not scraped)
- Kleijn, Bayesian Statistics for Frequentists, §6.4 (standard reference, not scraped)