Documentation

Exchangeability.Probability.TripleLawDropInfo.PairLawHelpers

Helper Lemmas for RN-Derivative Approach to Kallenberg 1.3 #

This file provides helper lemmas for the Radon-Nikodym derivative approach to Kallenberg Lemma 1.3 (contraction-independence).

Main results #

References #

theorem marginal_law_eq_of_pair_law {Ω : Type u_1} {α : Type u_2} {γ : Type u_3} [MeasurableSpace Ω] [MeasurableSpace α] [MeasurableSpace γ] {μ : MeasureTheory.Measure Ω} (X : Ωα) (W W' : Ωγ) (hX : Measurable X) (hW : Measurable W) (hW' : Measurable W') (h_law : MeasureTheory.Measure.map (fun (ω : Ω) => (X ω, W ω)) μ = MeasureTheory.Measure.map (fun (ω : Ω) => (X ω, W' ω)) μ) :

From pair law equality (X,W) =^d (X,W'), extract marginal law equality W =^d W'.

theorem joint_measure_eq_of_pair_law {Ω : Type u_1} {α : Type u_2} {γ : Type u_3} [MeasurableSpace Ω] [MeasurableSpace α] [MeasurableSpace γ] {μ : MeasureTheory.Measure Ω} (X : Ωα) (W W' : Ωγ) (hX : Measurable X) (hW : Measurable W) (hW' : Measurable W') (h_law : MeasureTheory.Measure.map (fun (ω : Ω) => (X ω, W ω)) μ = MeasureTheory.Measure.map (fun (ω : Ω) => (X ω, W' ω)) μ) {A : Set α} (hA : MeasurableSet A) :

From pair law equality, derive joint measure equality on the conditioning space.

If (X,W) =^d (X,W'), then map W (μ.restrict (X ⁻¹' A)) = map W' (μ.restrict (X ⁻¹' A)).

Intuitively: "the law of W restricted to {X ∈ A}" equals "the law of W' restricted to {X ∈ A}".

theorem integral_sq_condExp_eq_of_pair_law {Ω : Type u_1} {α : Type u_2} {γ : Type u_3} [MeasurableSpace Ω] [MeasurableSpace α] [MeasurableSpace γ] {μ : MeasureTheory.Measure Ω} [MeasureTheory.IsProbabilityMeasure μ] (X : Ωα) (W W' : Ωγ) (hX : Measurable X) (hW : Measurable W) (hW' : Measurable W') (h_law : MeasureTheory.Measure.map (fun (ω : Ω) => (X ω, W ω)) μ = MeasureTheory.Measure.map (fun (ω : Ω) => (X ω, W' ω)) μ) {A : Set α} (hA : MeasurableSet A) :
(ω : Ω), μ[(X ⁻¹' A).indicator fun (x : Ω) => 1|MeasurableSpace.comap W inferInstance] ω * μ[(X ⁻¹' A).indicator fun (x : Ω) => 1|MeasurableSpace.comap W inferInstance] ω μ = (ω : Ω), μ[(X ⁻¹' A).indicator fun (x : Ω) => 1|MeasurableSpace.comap W' inferInstance] ω * μ[(X ⁻¹' A).indicator fun (x : Ω) => 1|MeasurableSpace.comap W' inferInstance] ω μ

Helper for Kallenberg 1.3: Square integrals are equal via Doob-Dynkin factorization.

Given:

  • (X,W) =^d (X,W') (pair laws equal, hence ρ = law(W) = law(W'))
  • μ₁ = E[φ|σ(W)] and μ₂ = E[φ|σ(W')] where φ = 1_{X∈A}

This lemma proves ∫ μ₁² dμ = ∫ μ₂² dμ using:

  1. Doob-Dynkin: μ₁ = g₁ ∘ W, μ₂ = g₂ ∘ W' for measurable g₁, g₂
  2. Set-integral uniqueness: g₁ = g₂ ρ-a.e.
  3. Change of variables: ∫ μ₁² dμ = ∫ g₁² dρ = ∫ g₂² dρ = ∫ μ₂² dμ