(32), p. 352 — s(x) = ∂E[l(F+f)|x]/∂f(x) at f(x)=0 equals 2E(y* − p(x)|x)
ProvedAddLogReg.LogitBoost.eq_32boostinglogistic-regressionlogitboostnewton-methodp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1
Let be a probability measure on , the joint law of , let be the current fit and . Fix and consider the conditional expected log-likelihood of the update ,
Then is differentiable at with derivative
This is the score of the Newton step in the Derivation of Result 3.
Formalization Note The derivative is stated with HasDerivAt, so the statement includes differentiability. Conditional expectations are integrals against ν.condKernel x; the statement holds for every .
Preamble
import Mathlib import Definitions.Def_AddLogReg_LogitBoost_Setting open MeasureTheory ProbabilityTheory
Formal statement
namespace AddLogReg.LogitBoost
theorem eq_32 {X : Type*} [MeasurableSpace X] (ν : Measure (X × Bool)) [IsProbabilityMeasure ν]
(F : X → ℝ) (x : X) :
HasDerivAt (condLogLik ν F x)
(2 * ∫ b, (ystar b - symLogistic (F x)) ∂(ν.condKernel x)) 0 := by sorry
end AddLogReg.LogitBoost
Source
Friedman, Hastie and Tibshirani, Additive logistic regression: a statistical view of boosting, Ann. Statist. 28 (2000), p. 352, (32)
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.