Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

(36), p. 352 — the Newton update f(x) solves the weighted least-squares problem min E_w(F + ½z − (F + f))²

Proved
AddLogReg.LogitBoost.eq_36

by mikedeng1 · Oct 7, 2026 · Mathlib 0df444a (Lean v4.33.1)

boostinglogistic-regressionlogitboostp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1weighted-least-squares

Let ν\nuν be a probability measure on X×{0,1}X\times\{0,1\}X×{0,1}, F:X→RF : X\to\mathbb RF:X→R, p(x)=eF(x)/(eF(x)+e−F(x))p(x) = e^{F(x)}/(e^{F(x)}+e^{-F(x)})p(x)=eF(x)/(eF(x)+e−F(x)), z=(y∗−p(x))/(p(x)(1−p(x)))z = (y^*-p(x))/(p(x)(1-p(x)))z=(y∗−p(x))/(p(x)(1−p(x))) the working response and w(x)=p(x)(1−p(x))w(x) = p(x)(1-p(x))w(x)=p(x)(1−p(x)) the weight of Algorithm 3. Let

fN(x)=−H(x)−1s(x)=−2E(y∗−p(x)∣x)−4E(p(x)(1−p(x))∣x)f_N(x) = -H(x)^{-1}s(x) = -\frac{2E(y^*-p(x)\mid x)}{-4E(p(x)(1-p(x))\mid x)}fN​(x)=−H(x)−1s(x)=−−4E(p(x)(1−p(x))∣x)2E(y∗−p(x)∣x)​

be the Newton update of (34). Then fN(x)f_N(x)fN​(x) solves the weighted least-squares approximation about F(x)F(x)F(x) to the log-likelihood,

min⁡f(x)Ew(x)(F(x)+12 y∗−p(x)p(x)(1−p(x))−(F(x)+f(x)))2,\min_{f(x)} E_{w(x)}\Big(F(x) + \frac12\,\frac{y^*-p(x)}{p(x)(1-p(x))} - \big(F(x)+f(x)\big)\Big)^2,f(x)min​Ew(x)​(F(x)+21​p(x)(1−p(x))y∗−p(x)​−(F(x)+f(x)))2,

that is, for every real ttt,

Ew[(F(x)+12z−(F(x)+fN(x)))2 ∣ x]≤Ew[(F(x)+12z−(F(x)+t))2 ∣ x];E_w\Big[\big(F(x)+\tfrac12 z - (F(x)+f_N(x))\big)^2\,\Big|\,x\Big] \le E_w\Big[\big(F(x)+\tfrac12 z - (F(x)+t)\big)^2\,\Big|\,x\Big];Ew​[(F(x)+21​z−(F(x)+fN​(x)))2​x]≤Ew​[(F(x)+21​z−(F(x)+t))2​x];

and F(x)+fN(x)F(x)+f_N(x)F(x)+fN​(x) is the population LogitBoost update F(x)+12Ew(z∣x)F(x)+\frac12E_w(z\mid x)F(x)+21​Ew​(z∣x) of Algorithm 3.

Formalization Note The page writes the weighted expectation in (36) as Ew(x)E_{w(x)}Ew(x)​ without "∣x\mid x∣x"; since the Derivation conditions on xxx throughout and w(x)w(x)w(x) depends on xxx only, it is read as the conditional weighted expectation Ew(⋅∣x)E_w(\cdot\mid x)Ew​(⋅∣x).

Preamble
import Mathlib
import Definitions.Def_AddLogReg_LogitBoost_Setting

open MeasureTheory ProbabilityTheory
Formal statement
namespace AddLogReg.LogitBoost

theorem eq_36 {X : Type*} [MeasurableSpace X] (ν : Measure (X × Bool)) [IsProbabilityMeasure ν]
    (F : X → ℝ) (x : X) :
    (∀ t : ℝ,
      wCondExpBy ν (logitWeight F)
          (fun x' b => (F x' + (1 / 2) * workingResponse F x' b -
            (F x' + -((2 * ∫ b', (ystar b' - symLogistic (F x)) ∂(ν.condKernel x)) /
              (-4 * ∫ _b', symLogistic (F x) * (1 - symLogistic (F x)) ∂(ν.condKernel x))))) ^ 2) x ≤
        wCondExpBy ν (logitWeight F)
          (fun x' b => (F x' + (1 / 2) * workingResponse F x' b - (F x' + t)) ^ 2) x) ∧
      F x + -((2 * ∫ b, (ystar b - symLogistic (F x)) ∂(ν.condKernel x)) /
          (-4 * ∫ _b, symLogistic (F x) * (1 - symLogistic (F x)) ∂(ν.condKernel x))) =
        logitBoostStep ν F x := by sorry

end AddLogReg.LogitBoost
Source
Friedman, Hastie and Tibshirani, Additive logistic regression: a statistical view of boosting, Ann. Statist. 28 (2000), p. 352, (36)
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