Proposition 6, pp. 609–610 — nonnegative weights that reproduce bounded continuous satisfy (3)
ProvedStoneRegression.Criterion.proposition_6convergence-in-probabilitynonparametric-regressionp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1weight-functions
Let be i.i.d. -valued with law , and let be a sequence of nonnegative Borel weights such that, for every bounded continuous function on ,
Then satisfies condition (3):
This is the step from consistency to the localization condition (3) in the necessity half of Theorem 1 (applied to responses ).
Formalization Note. Convergence in probability is Mathlib's convergence in measure under the product law of ; bounded continuous functions are Mathlib's BoundedContinuousFunction. Weights are assumed jointly Borel (standing assumption).
Preamble
import Mathlib import Definitions.Def_StoneRegression_Criterion_Setting open MeasureTheory ProbabilityTheory Filter Topology open scoped ENNReal NNReal BoundedContinuousFunction
Formal statement
namespace StoneRegression.Criterion
/-- Proposition 6 (Stone 1977, pp. 609–610). Nonnegative weights such that `∑ᵢ W_{ni}(X) f(Xᵢ) → f(X)` in
probability for every bounded continuous `f` on `ℝᵈ` satisfy (3). -/
theorem proposition_6 {d : ℕ} (μ : Measure (EuclideanSpace ℝ (Fin d))) [IsProbabilityMeasure μ]
(W : WeightSeq d) (hW : MeasurableWeights W) (hnn : Nonneg W)
(hconv : ∀ f : EuclideanSpace ℝ (Fin d) →ᵇ ℝ, TendstoInMeasure (xLaw μ)
(fun n ω => ∑ i : Fin n, wAt W n ω i * f (ω (i.val + 1))) atTop (fun ω => f (ω 0))) :
Cond3 μ W := by sorry
end StoneRegression.Criterion
Source
Stone (1977), Ann. Statist. 5, Proposition 6, pp. 609–610
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.