condExp_comap_fst_eq_partial_integral
ProvedLet and be probability measures on measurable spaces and , and let be integrable on the product measure . Then the conditional expectation of given the sub--algebra generated by the FIRST coordinate (MeasurableSpace.comap Prod.fst) equals, -almost everywhere, the partial integral of over the second factor:
i.e. conditioning a product-measure integrand on the first factor amounts to integrating out the second factor. This is the mirror of the brick efron_stein_condExp_comap_snd_eq_partial_integral (which conditions on the second coordinate). It is a prerequisite for the multi-coordinate prefix-filtration conditional expectation used in the Doob-martingale proof of the Efron–Stein tensorization of variance. Source: R. van Handel, Probability in High Dimension (APC 550), §2.1; Boucheron–Lugosi–Massart, Concentration Inequalities (OUP 2013), Ch. 3.
import Mathlib.MeasureTheory.Function.ConditionalExpectation.Basic import Mathlib.MeasureTheory.Integral.Prod import Mathlib.MeasureTheory.Measure.Prod open MeasureTheory Filter open scoped ENNReal NNReal BigOperators
theorem condExp_comap_fst_eq_partial_integral
{β γ : Type*} [MeasurableSpace β] [MeasurableSpace γ]
(ρ : Measure β) [IsProbabilityMeasure ρ] (σ : Measure γ) [IsProbabilityMeasure σ]
{W : β × γ → ℝ} (hW : Integrable W (ρ.prod σ)) :
(ρ.prod σ)[W | MeasurableSpace.comap Prod.fst inferInstance]
=ᵐ[ρ.prod σ] fun p => ∫ y, W (p.1, y) ∂σ := by sorry