Lemma 4.2, p. 539 — |E^Y f(Y, Z) − E^Y f(Y, Z′)| ≤ 2‖f‖φ(m) for Y ∈ ℳ_{1k}, Z ∈ ℳ_{k+m,∞}
ProvedPoissonDepTrials.MixInv.lemma_4_2covariance-inequalitymixingp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1
Let be an arbitrary sequence of real random variables satisfying the mixing condition (4.1) with coefficients . Let be measurable with respect to and measurable with respect to , . Then for every bounded Borel function ,
where has the law of and is independent of the sequence. It is the decoupling inequality used in Lemmas 4.3 and 4.4.
Formalization Note The sequence is real valued and measurable (needed for to be a sub--algebra). is written . is written, almost surely, as , which avoids enlarging to carry . is a bound ; the lag satisfies , as in (4.1).
Preamble
import Mathlib import Definitions.Def_PoissonDepTrials_MixInv_Setting open MeasureTheory ProbabilityTheory
Formal statement
namespace PoissonDepTrials.MixInv
theorem lemma_4_2 {Ω : Type*} [MeasurableSpace Ω] (P : Measure Ω) [IsProbabilityMeasure P]
(X : ℕ → Ω → ℝ) (hXm : ∀ i, Measurable (X i)) (φ : ℕ → ℝ) (hφ : IbragimovMixing P X φ)
(r s k m : ℕ) (hm : 1 ≤ m)
(Y : Ω → (Fin r → ℝ)) (hY : Measurable[sigmaIcc X 1 k] Y)
(Z : Ω → (Fin s → ℝ)) (hZ : Measurable[sigmaIci X (k + m)] Z)
(f : (Fin r → ℝ) × (Fin s → ℝ) → ℝ) (hf : Measurable f) (M : ℝ) (hM : ∀ x, |f x| ≤ M) :
∀ᵐ ω ∂P, |(P[fun ω => f (Y ω, Z ω) | MeasurableSpace.comap Y inferInstance]) ω
- ∫ z, f (Y ω, z) ∂(P.map Z)| ≤ 2 * M * φ m := by sorry
end PoissonDepTrials.MixInv
Source
Chen, Poisson approximation for dependent trials, Ann. Probab. 3 (1975), p. 539, Lemma 4.2, (4.4)
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.