Documentation

Linglib.Core.Probability.Process.Stationary

Stationary processes #

A process X : ι → Ω → α is stationary when shifting time leaves its law on ι → α unchanged, which says that every shift of ι → α preserves that law. The finite-dimensional distributions determine the law, so stationarity can be checked on finite sets of times, and over ℕ on blocks of consecutive values, the form in which information theory defines it. Over ℕ a shift by one is enough, and a process that advances by a measure-preserving map of the sample space is stationary.

Implementation notes #

[UPSTREAM] candidate: mathlib has no notion of a stationary process.

References #

def ProbabilityTheory.IsStationary {ι : Type u_1} {Ω : Type u_2} {α : Type u_3} {mΩ : MeasurableSpace Ω} [MeasurableSpace α] [Add ι] (X : ι → Ω → α) (μ : MeasureTheory.Measure Ω) :

A process is (strictly) stationary when shifting time does not change its law.

Equations
  • ProbabilityTheory.IsStationary X μ = ∀ (s : ι), ProbabilityTheory.IdentDistrib (fun (ω : Ω) (t : ι) => X (t + s) ω) (fun (ω : Ω) (t : ι) => X t ω) μ μ
Instances For
    theorem ProbabilityTheory.IsStationary.comp {ι : Type u_1} {Ω : Type u_2} {α : Type u_3} {β : Type u_4} {mΩ : MeasurableSpace Ω} [MeasurableSpace α] [MeasurableSpace β] {μ : MeasureTheory.Measure Ω} [Add ι] {X : ι → Ω → α} (hs : IsStationary X μ) {f : α → β} (hf : Measurable f) :
    IsStationary (fun (t : ι) (ω : Ω) => f (X t ω)) μ
    theorem ProbabilityTheory.isStationary_iff_measurePreserving {ι : Type u_1} {Ω : Type u_2} {α : Type u_3} {mΩ : MeasurableSpace Ω} [MeasurableSpace α] {μ : MeasureTheory.Measure Ω} [Add ι] {X : ι → Ω → α} (hX : AEMeasurable (fun (ω : Ω) (t : ι) => X t ω) μ) :
    IsStationary X μ ↔ ∀ (s : ι), MeasureTheory.MeasurePreserving (fun (x : ι → α) (t : ι) => x (t + s)) (MeasureTheory.Measure.map (fun (ω : Ω) (t : ι) => X t ω) μ) (MeasureTheory.Measure.map (fun (ω : Ω) (t : ι) => X t ω) μ)
    theorem ProbabilityTheory.isStationary_iff_forall_finset {ι : Type u_1} {Ω : Type u_2} {α : Type u_3} {mΩ : MeasurableSpace Ω} [MeasurableSpace α] {μ : MeasureTheory.Measure Ω} [Add ι] {X : ι → Ω → α} [MeasureTheory.IsFiniteMeasure μ] (hX : AEMeasurable (fun (ω : Ω) (t : ι) => X t ω) μ) :
    IsStationary X μ ↔ ∀ (s : ι) (I : Finset ι), IdentDistrib (fun (ω : Ω) => I.restrict fun (t : ι) => X (t + s) ω) (fun (ω : Ω) => I.restrict fun (x : ι) => X x ω) μ μ
    def ProbabilityTheory.block {Ω : Type u_2} {α : Type u_3} (X : ℕ → Ω → α) (a n : ℕ) (ω : Ω) :
    Fin n → α

    block X a n is the block of n consecutive values of X starting at time a.

    Equations
    Instances For
      theorem ProbabilityTheory.block_zero_succ {Ω : Type u_2} {α : Type u_3} {X : ℕ → Ω → α} (t : ℕ) :
      block X 0 (t + 1) = ⇑(Fin.snocEquiv fun (x : Fin (t + 1)) => α) ∘ fun (ω : Ω) => (X t ω, block X 0 t ω)
      theorem ProbabilityTheory.block_zero_succ' {Ω : Type u_2} {α : Type u_3} {X : ℕ → Ω → α} (t : ℕ) :
      block X 0 (t + 1) = ⇑(Fin.consEquiv fun (x : Fin (t + 1)) => α) ∘ fun (ω : Ω) => (X 0 ω, block X 1 t ω)

      The block read from the front: its first value, then the block starting one step later.

      theorem ProbabilityTheory.measurable_block {Ω : Type u_2} {α : Type u_3} {mΩ : MeasurableSpace Ω} [MeasurableSpace α] {X : ℕ → Ω → α} (hX : ∀ (t : ℕ), Measurable (X t)) (a n : ℕ) :
      Measurable (block X a n)
      theorem ProbabilityTheory.IsStationary.identDistrib_block {Ω : Type u_2} {α : Type u_3} {mΩ : MeasurableSpace Ω} [MeasurableSpace α] {μ : MeasureTheory.Measure Ω} {X : ℕ → Ω → α} (hs : IsStationary X μ) (a n : ℕ) :
      IdentDistrib (block X a n) (block X 0 n) μ μ
      theorem ProbabilityTheory.isStationary_iff_forall_block {Ω : Type u_2} {α : Type u_3} {mΩ : MeasurableSpace Ω} [MeasurableSpace α] {μ : MeasureTheory.Measure Ω} {X : ℕ → Ω → α} [MeasureTheory.IsFiniteMeasure μ] :
      IsStationary X μ ↔ ∀ (a n : ℕ), IdentDistrib (block X a n) (block X 0 n) μ μ

      A process is stationary exactly when a block of n consecutive values has the same law wherever it starts, which is how information theory states the definition.

      theorem ProbabilityTheory.isStationary_iff_identDistrib_succ {Ω : Type u_2} {α : Type u_3} {mΩ : MeasurableSpace Ω} [MeasurableSpace α] {μ : MeasureTheory.Measure Ω} {X : ℕ → Ω → α} :
      IsStationary X μ ↔ IdentDistrib (fun (ω : Ω) (t : ℕ) => X (t + 1) ω) (fun (ω : Ω) (t : ℕ) => X t ω) μ μ
      theorem ProbabilityTheory.IsStationary.of_measurePreserving {Ω : Type u_2} {α : Type u_3} {mΩ : MeasurableSpace Ω} [MeasurableSpace α] {μ : MeasureTheory.Measure Ω} {X : ℕ → Ω → α} {T : Ω → Ω} (hT : MeasureTheory.MeasurePreserving T μ μ) (hX : ∀ (t : ℕ), Measurable (X t)) (h : ∀ (t : ℕ), X (t + 1) = X t ∘ T) :