Documentation

LeanPool.HighContrastHomogenization.Support.Book.Ch04.Theorems.RestrictionIndependence

Coarse-graining support: Support.Book.Ch04.Theorems.RestrictionIndependence #

Imported from the Apache-2.0 CoarseGraining development at commit c7ddd76c08ade64fed1b8d2ca51be14dfee8deb4.

Restriction-local independence and coloring lemmas #

This module owns the finite-family independence consequences of the explicit restriction-unit-range assumption. It concerns RestrictionSigmaR and whole-restriction-local random variables only; source-local finite independence belongs in a separate module.

Restriction-unit-range dependence gives independence of any pairwise separated finite family of carrier restriction σ-algebras.

theorem HCPolySupport.Book.Ch04.iIndepFun_of_restrictionUnitRangeDependentLaw_of_pairwise_separated {d : ℕ} {ι : Type u_1} {β : ι → Type u_2} [(i : ι) → MeasurableSpace (β i)] {P : RestrictionCoeffLaw d} [MeasureTheory.IsProbabilityMeasure P] {U : ι → Set (Vec d)} {X : (i : ι) → RegCoeffField d → β i} (hU : ∀ (i : ι), MeasurableSet (U i)) (hP : RestrictionUnitRangeDependentLaw P) (hX : ∀ (i : ι), IsRestrictionLocalRandomVariable (U i) ⋯ (X i)) (hsep : Pairwise fun (i j : ι) => AreUnitSeparated (U i) (U j)) :

Restriction-local random variables indexed by pairwise separated observation sets are independent under restriction-unit-range dependence.

A single scale-color class of descendant cube observables is an independent family under the explicit restriction-unit-range dependence assumption.