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 entropy for compact-location normal mixtures in a scale band
Statement
For , , and , let There is a universal constant such that The same bound, with a changed universal constant, holds for whenever for a sufficiently small universal . Here is the covering number in the stated metric.
Facts & Assumptions
Given: , , , AC, and the mixture density of The Dirichlet normal-mixture density is jointly measurable and normalized.
Assume AC for the normal-density and finite-support approximation interfaces. The Axiom of Choice
Compactly supported Gaussian mixtures admit logarithmic-size finite approximations: a mixture with mixing law on is within error of a discrete mixture with at most atoms.
Hellinger distance is a metric and controls L1 and total variation: for probability densities, Hellinger distance is a metric and .
A positive-variance normal law has its location-scale density: , with the standard normal density.
Root-L2 Hellinger distance between density laws: the Hellinger metric used here is the root-L2 distance between square-root densities.
For a standard normal random variable , . Location-scale normal laws have their stated mean and variance
Nonnegative product integrals may be interchanged. Tonelli's theorem for nonnegative measurable functions on a sigma-finite product
A differentiable function with integrable derivative on a compact interval has endpoint difference equal to the integral of its derivative. The second fundamental theorem: if is differentiable on with and is integrable, then
Proof
Given: The hypotheses in the Statement.
Put . By [F1], every is within distance of a mixture with at most atoms, for a fixed sufficiently large . Pad shorter mixtures by zero-weight atoms.
Differentiating the density in [F3] gives , since and by [F5], and for . Integrating these derivatives along the location and scale segments by [F7] yields and . For the first class use a scale grid in ; for an actual band, instead grid , omitting when that endpoint is excluded. In either case choose grid mesh at most , with every grid point admitted by the corresponding scale constraint. Use a location grid in with mesh at most and an grid on the -simplex of mesh at most . Replacing the scale, locations and weights changes the mixture density in by at most : integrate the componentwise location and scale bounds against the finite mixing weights, and use Tonelli [F6] for the componentwise integrals. Choose so this plus the approximation from step 1.1 is at most . Thus the resulting finite set, with centers in the class being covered, is an cover of radius , hence an -cover of radius by [F2].
The scale grid has at most points and the location grid at most points. Rounding each of the first weights down to a mesh proportional to leaves the last weight nonnegative after setting it to make the sum one, gives at most weight vectors, and changes their distance by at most . Therefore the logarithm of the cover size is at most The constructed finite set is an -net of radius , hence also a net of radius ; its cardinality therefore bounds the covering number and proves the first bound.
If , then already belongs to the compact-location class. If , let and be the normalized restrictions to and its complement. Then , so by integrating the mixture densities (Tonelli [F6]). By [F2], ; the triangle inequality in [F4] then transfers the cover to the tail-constrained class. The scale grid is restricted as in step 2.1, so every center has and satisfies the band and cutoff constraints. For sufficiently small universal , makes the truncation error at most , so the radius- cover from steps 2.1 and 3.1 becomes a radius- cover after the constant rescaling used in the entropy bound.
Depends on
- The Dirichlet normal-mixture density is jointly measurable and normalized
- Compactly supported Gaussian mixtures admit logarithmic-size finite approximations
- Root-L2 Hellinger distance between density laws
- Hellinger distance is a metric and controls L1 and total variation
- A positive-variance normal law has its location-scale density
- The Axiom of Choice
- Location-scale normal laws have their stated mean and variance
- Tonelli's theorem for nonnegative measurable functions on a sigma-finite product
- The second fundamental theorem: if $G$ is differentiable on $[a,b]$ with $G' = f$ and $f$ is integrable, then $\int_a^b f = G(b)-G(a)$
Used by
Dependency tree · two levels
57 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
- Subhashis Ghosal and Aad W. van der Vaart, Posterior convergence rates of Dirichlet mixtures at smooth densities, §5 Theorem 6 (standard reference, not scraped)
- Subhashis Ghosal and Aad W. van der Vaart, Fundamentals of Nonparametric Bayesian Inference, §6.3, proof of Theorem 6.6 (standard reference, not scraped)