(4.7), p. 540 — Var(W) = λ + ΣΣ_{0<|i−j|≤m} EX_iX_j − ΣΣ_{|i−j|≤m} p_ip_j + Σ_i E[(X_i − p_i)V^{(i)}]
ProvedPoissonDepTrials.MixSqrt.variance_identity_4_7bernoulli-trialsp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1poisson-approximationvariance
Let be Bernoulli random variables on a probability space , possibly dependent, with , and put and . As on p. 535, is taken to be identically zero when or . Fix . Then
No assumption on the dependence is made; the identity is (2.2) with .
It is the first step of the proofs of Lemmas 4.4 and 4.5: it expresses the local correlation sum through , which is what lets the final bound be stated in terms of .
Formalization Note The trials are a sequence that is measurable, almost surely -valued, and identically at index and above (the paper's padding convention, which changes no -algebra). is defined as , not a free parameter.
Preamble
import Mathlib import Definitions.Def_PoissonDepTrials_MixSqrt_Setting open MeasureTheory ProbabilityTheory
Formal statement
namespace PoissonDepTrials.MixSqrt
/-- Chen (1975), proof of Lemma 4.4, p. 540, (4.7): for an arbitrary sequence of dependent
Bernoulli trials,
`Var(W) = λ + ΣΣ_{0<|i−j|≤m} EX_iX_j − ΣΣ_{|i−j|≤m} p_ip_j + Σ_i E[(X_i − p_i)V^{(i)}]`. -/
theorem variance_identity_4_7 {Ω : Type*} [MeasurableSpace Ω] (P : Measure Ω) [IsProbabilityMeasure P]
(n : ℕ) (X : ℕ → Ω → ℕ) (hX : IsBernoulliTrials P n X) (m : ℕ) :
Var[fun ω => (W n X ω : ℝ); P] = lam P n X
+ (∑ i ∈ Finset.Icc 1 n, ∑ j ∈ Finset.Icc 1 n with 0 < |(i : ℤ) - (j : ℕ)| ∧ |(i : ℤ) - (j : ℕ)| ≤ m,
∫ ω, (X i ω : ℝ) * X j ω ∂P)
- (∑ i ∈ Finset.Icc 1 n, ∑ j ∈ Finset.Icc 1 n with |(i : ℤ) - (j : ℕ)| ≤ m,
prob P X i * prob P X j)
+ ∑ i ∈ Finset.Icc 1 n, ∫ ω, ((X i ω : ℝ) - prob P X i) * (V n m X i ω : ℝ) ∂P := by sorry
end PoissonDepTrials.MixSqrt
Source
Chen, Poisson approximation for dependent trials, Ann. Probab. 3 (1975), p. 540, proof of Lemma 4.4, (4.7)
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.