Radon-Nikodym theorem #
This file proves the Radon-Nikodym theorem. The Radon-Nikodym theorem states that, given measures
μ, ν, if HaveLebesgueDecomposition μ ν, then μ is absolutely continuous with respect to
ν if and only if there exists a measurable function f : α → ℝ≥0∞ such that μ = fν.
In particular, we have f = rnDeriv μ ν.
The Radon-Nikodym theorem will allow us to define many important concepts in probability theory,
most notably probability cumulative functions. It could also be used to define the conditional
expectation of a real function, but we take a different approach (see the file
MeasureTheory/Function/ConditionalExpectation).
Main results #
- MeasureTheory.Measure.absolutelyContinuous_iff_withDensity_rnDeriv_eq: the Radon-Nikodym theorem
- MeasureTheory.SignedMeasure.absolutelyContinuous_iff_withDensityᵥ_rnDeriv_eq: the Radon-Nikodym theorem for signed measures
The file also contains properties of rnDeriv that use the Radon-Nikodym theorem, notably
- MeasureTheory.Measure.rnDeriv_withDensity_left: the Radon-Nikodym derivative of- μ.withDensity fwith respect to- νis- f * μ.rnDeriv ν.
- MeasureTheory.Measure.rnDeriv_withDensity_right: the Radon-Nikodym derivative of- μwith respect to- ν.withDensity fis- f⁻¹ * μ.rnDeriv ν.
- MeasureTheory.Measure.inv_rnDeriv:- (μ.rnDeriv ν)⁻¹ =ᵐ[μ] ν.rnDeriv μ.
- MeasureTheory.Measure.setLIntegral_rnDeriv:- ∫⁻ x in s, μ.rnDeriv ν x ∂ν = μ sif- μ ≪ ν. There is also a version of this result for the Bochner integral.
Tags #
Radon-Nikodym theorem
The Radon-Nikodym theorem: Given two measures μ and ν, if
HaveLebesgueDecomposition μ ν, then μ is absolutely continuous to ν if and only if
ν.withDensity (rnDeriv μ ν) = μ.
Auxiliary lemma for rnDeriv_withDensity_left.
Auxiliary lemma for rnDeriv_withDensity_right.
Auxiliary lemma for rnDeriv_withDensity_right.
Auxiliary lemma for rnDeriv_add_right_of_mutuallySingular.
Auxiliary lemma for rnDeriv_add_right_of_mutuallySingular.
Auxiliary lemma for inv_rnDeriv.
See also setIntegral_rnDeriv_smul' for a version that requires both measures to be σ-finite,
but doesn't require s to be a measurable set.
A version of setIntegral_rnDeriv_smul that requires both measures to be σ-finite,
but doesn't require s to be a measurable set.