Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

§2, display (6), p. 364 — the stochastic kriging predictor is the unique MSE-optimal predictor of the form (5)

Proved
StochKriging.OptimalMSE.display_6

by mikedeng1 · Oct 6, 2026 · Mathlib 0df444a (Lean v4.33.1)

best-linear-predictorkrigingmean-squared-errorp2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-paperp2o-v1

Work in the second-order stochastic kriging model: outputs Yj(x)=β0+M(x)+εj(x)\mathcal Y_j(\mathbf x)=\beta_0+\mathsf M(\mathbf x)+\varepsilon_j(\mathbf x)Yj​(x)=β0​+M(x)+εj​(x) with square-integrable, mean-zero field M\mathsf MM and noise ε\varepsilonε, field and noise uncorrelated, and noise correlation across design points allowed (CRN). Let (xi,ni)i=1k(\mathbf x_i,n_i)_{i=1}^k(xi​,ni​)i=1k​ be a design with every ni≥1n_i\ge1ni​≥1, let Yˉ\bar{\mathcal Y}Yˉ​ be the vector of sample means, and let ΣM\Sigma_{\mathsf M}ΣM​, Σε\Sigma_\varepsilonΣε​ and ΣM(x0,⋅)\Sigma_{\mathsf M}(\mathbf x_0,\cdot)ΣM​(x0​,⋅) be the covariance matrices and vector of §2. Assume ΣM+Σε\Sigma_{\mathsf M}+\Sigma_\varepsilonΣM​+Σε​ is positive definite, and put

λ⋆=[ΣM+Σε]−1ΣM(x0,⋅),λ0⋆=β0−β0 λ⋆⊤1k.\lambda^\star=[\Sigma_{\mathsf M}+\Sigma_\varepsilon]^{-1}\Sigma_{\mathsf M}(\mathbf x_0,\cdot),\qquad \lambda_0^\star=\beta_0-\beta_0\,\lambda^{\star\top}\mathbf 1_k .λ⋆=[ΣM​+Σε​]−1ΣM​(x0​,⋅),λ0⋆​=β0​−β0​λ⋆⊤1k​.

Then, for every prediction point x0\mathbf x_0x0​:

  1. the stochastic kriging predictor Y^(x0)=β0+ΣM(x0,⋅)⊤[ΣM+Σε]−1(Yˉ−β01k)\widehat{\mathsf Y}(\mathbf x_0)=\beta_0+\Sigma_{\mathsf M}(\mathbf x_0,\cdot)^\top[\Sigma_{\mathsf M}+\Sigma_\varepsilon]^{-1}(\bar{\mathcal Y}-\beta_0\mathbf 1_k)Y(x0​)=β0​+ΣM​(x0​,⋅)⊤[ΣM​+Σε​]−1(Yˉ​−β0​1k​) is the linear predictor λ0⋆+λ⋆⊤Yˉ\lambda_0^\star+\lambda^{\star\top}\bar{\mathcal Y}λ0⋆​+λ⋆⊤Yˉ​ of the form (5);
  2. it has the smallest mean squared error for Y(x0)=β0+M(x0)\mathsf Y(\mathbf x_0)=\beta_0+\mathsf M(\mathbf x_0)Y(x0​)=β0​+M(x0​) among all predictors of the form (5):
E[(Y^(x0)−Y(x0))2]≤E[(λ0+λ⊤Yˉ−Y(x0))2]for all λ0∈R, λ∈Rk;\mathrm E\big[(\widehat{\mathsf Y}(\mathbf x_0)-\mathsf Y(\mathbf x_0))^2\big]\le\mathrm E\big[(\lambda_0+\lambda^\top\bar{\mathcal Y}-\mathsf Y(\mathbf x_0))^2\big]\quad\text{for all }\lambda_0\in\mathbb R,\ \lambda\in\mathbb R^k;E[(Y(x0​)−Y(x0​))2]≤E[(λ0​+λ⊤Yˉ​−Y(x0​))2]for all λ0​∈R, λ∈Rk;
  1. it is the only one: any weights (λ0,λ)(\lambda_0,\lambda)(λ0​,λ) achieving an MSE no larger than that of Y^(x0)\widehat{\mathsf Y}(\mathbf x_0)Y(x0​) equal (λ0⋆,λ⋆)(\lambda_0^\star,\lambda^\star)(λ0⋆​,λ⋆).

This is the paper's display (6), which defines stochastic kriging: the classical kriging predictor with the covariance of the data enlarged by the intrinsic covariance Σε\Sigma_\varepsilonΣε​.

Formalization Note The weights range over all of R×Rk\mathbb R\times\mathbb R^kR×Rk, with no unbiasedness constraint. No Gaussianity, no independence across replications and no "no CRN" assumption is made; Σε\Sigma_\varepsilonΣε​ is a general covariance matrix. The uncorrelatedness of field and noise and the finite second moments are added (§2 uses them implicitly). Positive definiteness of ΣM+Σε\Sigma_{\mathsf M}+\Sigma_\varepsilonΣM​+Σε​ makes the inverse in (6) genuine; the sum is always positive semidefinite, being the covariance matrix of Yˉ\bar{\mathcal Y}Yˉ​, so the hypothesis says exactly that it is invertible.

Preamble
import Mathlib
import Definitions.Def_StochKriging_OptimalMSE_Model

open MeasureTheory ProbabilityTheory Matrix
Formal statement
namespace StochKriging.OptimalMSE

/-- Display (6), p. 364: the stochastic kriging predictor `Ŷ(x₀)` is a predictor of the form (5),
with weights `λ = [ΣM + Σε]⁻¹ ΣM(x₀, ·)` and `λ₀ = β₀ − β₀ λᵀ1_k`; it has the smallest MSE for
`Y(x₀)` among all predictors of the form (5); and these weights are the only minimizers. -/
theorem display_6 {Ω X : Type*} [MeasurableSpace Ω] {P : Measure Ω} [IsProbabilityMeasure P]
    {k : ℕ} (β₀ : ℝ) (M : X → Ω → ℝ) (ε : ℕ → X → Ω → ℝ) (x : Fin k → X) (n : Fin k → ℕ)
    (hmodel : IsModel P M ε) (hn : ∀ i, 1 ≤ n i)
    (hSig : (SigmaM P M x + SigmaEps P ε x n).PosDef) (x₀ : X) :
    (∀ ω, skPredictor P β₀ M ε x n x₀ ω =
        linPred
          (β₀ - β₀ * (((SigmaM P M x + SigmaEps P ε x n)⁻¹ *ᵥ SigmaMCross P M x x₀) ⬝ᵥ
            fun _ => 1))
          ((SigmaM P M x + SigmaEps P ε x n)⁻¹ *ᵥ SigmaMCross P M x x₀)
          (sampleMean β₀ M ε x n) ω) ∧
    (∀ (l₀ : ℝ) (l : Fin k → ℝ),
        mse P (skPredictor P β₀ M ε x n x₀) (target β₀ M x₀) ≤
          mse P (linPred l₀ l (sampleMean β₀ M ε x n)) (target β₀ M x₀)) ∧
    (∀ (l₀ : ℝ) (l : Fin k → ℝ),
        mse P (linPred l₀ l (sampleMean β₀ M ε x n)) (target β₀ M x₀) ≤
            mse P (skPredictor P β₀ M ε x n x₀) (target β₀ M x₀) →
          l = (SigmaM P M x + SigmaEps P ε x n)⁻¹ *ᵥ SigmaMCross P M x x₀ ∧
          l₀ = β₀ - β₀ * (((SigmaM P M x + SigmaEps P ε x n)⁻¹ *ᵥ SigmaMCross P M x x₀) ⬝ᵥ
            fun _ => 1)) := by sorry

end StochKriging.OptimalMSE
Source
Ankenman, Nelson, Staum, Stochastic Kriging for Simulation Metamodeling, Proc. 2008 Winter Simulation Conference, p. 364, §2, display (6)
Human review
  • Endorsed by Shuze Chen · Oct 7, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 7, 2026

    Confirmed by the mission captain (proposal self-audit).

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me