Documentation

Verification.FairSign

← Mathematical handbook

Uniform ranks from a uniform magnitude and a fair sign #

The sign selector may depend on the magnitude and on arbitrary other data. Symmetrizing it with its opposite still produces the uniform rank law.

noncomputable def Verification.fairMeasure {Ω : Type u_1} {X : Type u_2} [MeasurableSpace Ω] [MeasurableSpace X] (μ : MeasureTheory.Measure Ω) (f g : Ω → X) :
Equations
Instances For
    theorem Verification.integral_fairMeasure {Ω : Type u_1} {X : Type u_2} [MeasurableSpace Ω] [MeasurableSpace X] (μ : MeasureTheory.Measure Ω) {f g : Ω → X} {h : X → ℝ} (hf : Measurable f) (hg : Measurable g) (hh : Measurable h) (hi : MeasureTheory.Integrable h (MeasureTheory.Measure.map f μ)) (hj : MeasureTheory.Integrable h (MeasureTheory.Measure.map g μ)) :
    ∫ (x : X), h x ∂fairMeasure μ f g = 1 / 2 * ∫ (x : Ω), h (f x) ∂μ + 1 / 2 * ∫ (x : Ω), h (g x) ∂μ
    theorem Verification.fairMeasure_map {Ω : Type u_1} {X : Type u_2} {Y : Type u_3} [MeasurableSpace Ω] [MeasurableSpace X] [MeasurableSpace Y] (μ : MeasureTheory.Measure Ω) {f g : Ω → X} {h : X → Y} (hf : Measurable f) (hg : Measurable g) (hh : Measurable h) :
    theorem Verification.measurable_signedRank {Ω : Type u_1} [MeasurableSpace Ω] {A : Ω → ↑unitInterval} {S : Ω → Bool} (hA : Measurable A) (hS : Measurable S) :
    Measurable fun (w : Ω) => signedRank (S w) (A w)
    theorem Verification.fair_signed_rank_uniform {Ω : Type u_1} [MeasurableSpace Ω] (μ : MeasureTheory.Measure Ω) [MeasureTheory.IsProbabilityMeasure μ] {A : Ω → ↑unitInterval} {S : Ω → Bool} (hA : Measurable A) (hS : Measurable S) (hU : MeasureTheory.Measure.map A μ = MeasureTheory.volume) :
    (fairMeasure μ (fun (w : Ω) => signedRank (S w) (A w)) fun (w : Ω) => signedRank (!S w) (A w)) = MeasureTheory.volume