Lemma B.1 (20), p. 2325 — E[φ₁(ξ₁)φ₂(ξ₂)] ≤ E[φ₁(ξ₁)φ₂(ξ₁)] + E[φ₁(ξ₂)φ₂(ξ₂)] without independence
ProvedHighDimCLT.Hyperrect.lemma_B_1_eq_20association-inequalityhigh-dimensional-cltp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1
Let be nondecreasing functions and let be real-valued random variables, not necessarily independent. Assume that , and have finite expectations. Then
This is the last sentence of Lemma B.1 ("(20) holds without independence"), a variant of Chebyshev's association inequality; it is the form used in the proof of Lemma 5.1.
Formalization Note The random variables are measurable functions on a probability space; the three integrability hypotheses are the page's "all the expectations exist and are finite" for the expectations in (20).
Preamble
import Mathlib import Definitions.Def_HighDimCLT_Hyperrect_Setting open MeasureTheory ProbabilityTheory
Formal statement
namespace HighDimCLT.Hyperrect
theorem lemma_B_1_eq_20 {Ω : Type*} [MeasurableSpace Ω] (P : Measure Ω) [IsProbabilityMeasure P]
(φ₁ φ₂ : ℝ → ℝ) (hφ₁ : Monotone φ₁) (hφ₂ : Monotone φ₂)
(hφ₁0 : ∀ x, 0 ≤ φ₁ x) (hφ₂0 : ∀ x, 0 ≤ φ₂ x)
(ξ₁ ξ₂ : Ω → ℝ) (hξ₁ : Measurable ξ₁) (hξ₂ : Measurable ξ₂)
(hi₁₁ : Integrable (fun ω => φ₁ (ξ₁ ω) * φ₂ (ξ₁ ω)) P)
(hi₂₂ : Integrable (fun ω => φ₁ (ξ₂ ω) * φ₂ (ξ₂ ω)) P)
(hi₁₂ : Integrable (fun ω => φ₁ (ξ₁ ω) * φ₂ (ξ₂ ω)) P) :
∫ ω, φ₁ (ξ₁ ω) * φ₂ (ξ₂ ω) ∂P ≤
(∫ ω, φ₁ (ξ₁ ω) * φ₂ (ξ₁ ω) ∂P) + ∫ ω, φ₁ (ξ₂ ω) * φ₂ (ξ₂ ω) ∂P := by sorry
end HighDimCLT.Hyperrect
Source
Chernozhukov, Chetverikov and Kato, Central limit theorems and bootstrap in high dimensions, Ann. Probab. 45 (2017), pp. 2324–2325, App. B, Lemma B.1, display (20) and last sentence
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.