Documentation

TauCeti.MeasureTheory.Measure.Atom

Images of measure atoms #

A positive finite-mass measurable atom is a measurable set whose measurable subsets have either zero or full mass. Every almost-everywhere measurable map sends such an atom to one point of a standard Borel target: the image of the measure restricted to the atom is a scalar multiple of a Dirac measure.

The finite-mass hypothesis is essential: a zero-infinity measure may make its whole carrier an atom while vanishing on every singleton.

Main definitions #

Main results #

def MeasureTheory.Measure.IsAtom {X : Type u_1} [MeasurableSpace X] (μ : Measure X) (A : Set X) :

A measure atom is a measurable positive-mass set whose measurable subsets have either zero measure or the full measure of the set.

Equations
Instances For
    @[simp]
    theorem MeasureTheory.Measure.isAtom_iff {X : Type u_1} [MeasurableSpace X] {μ : Measure X} {A : Set X} :
    μ.IsAtom A ↔ MeasurableSet A ∧ 0 < μ A ∧ ∀ ⦃B : Set X⦄, MeasurableSet B → B ⊆ A → μ B = 0 ∨ μ B = μ A

    The defining characterization of a measure atom.

    theorem AEMeasurable.exists_ae_eq_const_restrict_of_atom {X : Type u_1} {Y : Type u_2} [MeasurableSpace X] [MeasurableSpace Y] [StandardBorelSpace Y] {T : X → Y} {μ : MeasureTheory.Measure X} {A : Set X} (hT : AEMeasurable T (μ.restrict A)) (hAfin : μ A ≠ ⊤) (hAatom : μ.IsAtom A) :
    ∃ (y : Y), T =ᵐ[μ.restrict A] fun (x : X) => y

    An a.e.-measurable map is a.e. constant on a measurable atom. If A has positive finite mass and every measurable subset of A has either zero or full mass, then a map from A into a standard Borel space agrees almost everywhere with a constant.

    theorem AEMeasurable.exists_map_restrict_eq_smul_dirac_of_atom {X : Type u_1} {Y : Type u_2} [MeasurableSpace X] [MeasurableSpace Y] [StandardBorelSpace Y] {T : X → Y} {μ : MeasureTheory.Measure X} {A : Set X} (hT : AEMeasurable T (μ.restrict A)) (hAfin : μ A ≠ ⊤) (hAatom : μ.IsAtom A) :

    An a.e.-measurable map sends a measurable atom to a point mass. If A has positive finite mass and every measurable subset of A has either zero or full mass, then the image of μ.restrict A under a map to a standard Borel space is μ A times a Dirac measure.

    theorem MeasureTheory.Measure.exists_eq_smul_dirac_of_forall_restrict_eq_smul {Z : Type u_3} [MeasurableSpace Z] [StandardBorelSpace Z] [Nonempty Z] (ν : Measure Z) [IsFiniteMeasure ν] (h : ∀ (s : Set Z), MeasurableSet s → ∃ (c : ENNReal), ν.restrict s = c • ν) :
    ∃ (z : Z), ν = ν Set.univ • dirac z

    A finite measure proportional to each of its restrictions is a point mass. On a standard Borel space, if the restriction of a finite measure ν to every measurable set is a scalar multiple of ν, then ν is its total mass times a Dirac measure.

    This identifies the measures spanning extreme rays of the cone of finite measures as the positive multiples of point masses; it is used to show that the extreme rays of the completely monotone cone are exponentials.

    An injective measurable map into a space whose singletons are measurable sends a measure with null singletons to a measure with null singletons: each singleton has an at-most-singleton preimage, which is null.

    Adapted from Cameron Freer's private noAtoms_map_of_injective in Graphon/MeasureIso.lean at commit 9f7be59fa754d260a544b4cfd83d6a5b94f7552e: https://github.com/cameronfreer/graphon/commit/9f7be59fa754d260a544b4cfd83d6a5b94f7552e; the original work is copyright Cameron Freer and licensed under Apache 2.0, and it assumes f to be a measurable embedding, where the measurability and injectivity of f are separated here.