Alphabeta Math
TheoremStatement: AI-adaptedProof: AI-adaptedPipeline-generatedjudge pass (gpt-6-luna)
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.

Vanishing bias of a smoothed regular-grid spatial periodogram

Statement

Assume the Axiom of Choice and Dependent Choice, as in Fejer convolution formula for the lattice periodogram. Let the lattice spectral measure from that theorem be absolutely continuous with a specified bounded representative f∈L∞(Td) continuous at the chosen target frequency λ0∈Td. Let K and f~n,b be as in Regular-grid spatial periodogram and smoothed estimator, and let bn>0 satisfy bn→0,nbn→∞. Then the denominator defining f~n,bn(λ0) is positive for all sufficiently large n, and E[In(λ0)]⟶f(λ0),E[f~n,bn(λ0)]⟶f(λ0). This is an asymptotic-bias statement only; it asserts no variance limit, weak consistency, or central limit theorem.

Facts & Assumptions

Given: The known-mean regular-grid periodogram, periodized kernel Kbper, bandwidth sequence, and continuous bounded spectral density in the Statement. For angular representatives define ρT(x,y):=min⁡m∈Z∣x−y+2πm∣,ρ∞(λ,θ):=max⁡1≤r≤dρT(λr,θr), and for u∈Rd define ∥u∥∞:=max⁡r∣ur∣.

[A1]

Regular-grid spatial periodogram and smoothed estimator: the normalized Fourier-grid weights are wn,q,b(λ) and f~n,b(λ)=∑qwn,q,b(λ)In(λn,q) when the denominator is positive.

[A2]

The Axiom of Choice: the Axiom of Choice is assumed for the imported Bochner representation.

[A3]

The axiom of dependent choice: a relation in which every element is related to something admits an N-indexed chain: Dependent Choice is also assumed for that representation, as in the preceding Fejer theorem.

[F1]

Fejer convolution formula for the lattice periodogram: the expected raw periodogram is gn(λ):=E[In(λ)]=∫TdΦn(λ−θ)f(θ) dθ, where Φn≥0 and ∫TdΦn=1.

[F2]

Signs, monotonicity intervals, and ranges of sine and cosine and Pi is the first positive zero of sine: sine is increasing on [0,π/2] and positive on (0,π), so sin⁡(δ/2)>0 for 0<δ<π.

[F3]

Linearity, monotonicity, and the modulus bound for expectation: expectation passes through the finite weighted sum defining the smoother.

[F4]

The product measure of two sigma-finite measure spaces and Fubini's theorem for L^1 functions on a sigma-finite product: product angular measure and iterated integrals for the product Fejer kernel use the stated Lebesgue product measure.

Proof

Proof technique: control the Fejer mass outside neighborhoods, then use the kernel support and Fourier-grid mesh to localize the smoothing weights.

1.1A2A3F1F2F4givenalgebra

Write ϕn(x):=(2πn)−1∣∑a=0n−1e−iax∣2, so Φn(x)=∏r=1dϕn(xr) by [F1]. Fix 0<δ<π. If the circular distance of x from 0 is at least δ, choose its representative in [−π,π]; then ∣sin⁡(x/2)∣≥sin⁡(δ/2)>0 by [F2]. The finite geometric-series formula therefore gives ϕn(x)=12πn∣sin⁡(nx/2)sin⁡(x/2)∣2≤12πnsin⁡2(δ/2). By the union bound over coordinates and the unit integral of each factor from [F1], ∫{x:ρ∞(x,0)≥δ}Φn(x) dx≤dnsin⁡2(δ/2), where ρ∞ is the maximum circular distance over the d coordinates.

2.1A2A3F1step 1.1given

Fix ε>0. Continuity of f at λ0 gives 0<δ<π such that ∣f(θ)−f(λ0)∣<ε whenever ρ∞(θ,λ0)<2δ. For any λ′ with ρ∞(λ′,λ0)<δ, split the convolution in [F1] into that neighborhood and its complement. On the neighborhood its contribution to ∣gn(λ′)−f(λ0)∣ is at most ε because Φn has integral one. On the complement, ρ∞(λ′−θ,0)≥δ; with M:=∥f∥∞+∣f(λ0)∣<∞, step 1.1 gives ∣gn(λ′)−f(λ0)∣≤ε+Mdnsin⁡2(δ/2). This bound holds uniformly for all such λ′.

3.1A1givenstep 2.1algebra

Since K is continuous and K(0)>0, there is a>0 such that K(u)>0 whenever ∥u∥∞<a. A nearest Fourier-grid point to λ0 is within π/n in each circular coordinate. Since nbn→∞, eventually π/(nbn)<a, so the periodized kernel is positive at that grid point and the denominator in [A1] is positive. Moreover, if a weight in [A1] is positive, compact support of K implies its grid point is within circular distance Rbn of λ0 in every coordinate, where [−R,R]d contains the support of K. Since bn→0, all positive-weight grid points eventually lie within the δ-neighborhood from step 2.1.

4.1A1A2A3F3step 2.1step 3.1algebra∎

For all sufficiently large n, [F3] and [A1] give E[f~n,bn(λ0)]=∑q∈Λnwn,q,bn(λ0)gn(λn,q). The weights are nonnegative and sum to one, and every positive weight is on a grid point covered by step 2.1. Hence ∣E[f~n,bn(λ0)]−f(λ0)∣≤ε+Mdnsin⁡2(δ/2). Taking n→∞ and then arbitrary ε↓0 proves the smoothed bias limit. Taking λ′=λ0 in step 2.1 proves the raw expectation limit as well.

Depends on

Used by

Cited to discharge well-definedness by Regular-grid spatial periodogram and smoothed estimator.

Dependency tree · two levels

51 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