InformationTheory

InformationTheory.Shannon.SMB.AlgoetCover.Boundedness

source

D.6 — Boundedness (hoisted before D.5 because the liminf transfer uses #

blockLogAvg_bddAbove_ae to establish μZ-a.s. upper boundedness of blockLogAvgZ).

theorem

InformationTheory.Shannon.blockLogAvg_bddAbove_ae

source
{Ω : Type u_1} [MeasurableSpace Ω] {α : Type u_2} [Fintype α] [Nonempty α] [MeasurableSpace α] [MeasurableSingletonClass α] (μ : MeasureTheory.Measure Ω) [MeasureTheory.IsProbabilityMeasure μ] (p : ErgodicProcess μ α) :
∀ᵐ (ω : Ω) μ, Filter.IsBoundedUnder (fun (x1 x2 : ) => x1 x2) Filter.atTop fun (n : ) => blockLogAvg μ p.toStationaryProcess n ω

A.s. boundedness above for blockLogAvg.

A.s., blockLogAvg ≤ negLogQk(k=0)/n + 2·log n / n (from blockLogAvg_le_negLogQk_plus_error), and the RHS converges a.s. to conditionalEntropyTail μ p 0 (finite), hence the RHS is eventually bounded above and so is blockLogAvg.

Used by
    theorem

    InformationTheory.Shannon.blockLogAvg_bddBelow_ae

    source
    {Ω : Type u_1} [MeasurableSpace Ω] {α : Type u_2} [MeasurableSpace α] (μ : MeasureTheory.Measure Ω) [MeasureTheory.IsProbabilityMeasure μ] (p : ErgodicProcess μ α) :
    ∀ᵐ (ω : Ω) μ, Filter.IsBoundedUnder (fun (x1 x2 : ) => x1 x2) Filter.atTop fun (n : ) => blockLogAvg μ p.toStationaryProcess n ω

    A.s. boundedness below for blockLogAvg.

    Used by