(38), p. 354 — the LogitBoost increment of (34) equals ½ (P − p(x))/(p(x)(1 − p(x)))
ProvedAddLogReg.Gentle.eq_38boostinggentle-adaboostlogitboostp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1
Let be a probability measure on , the joint law of , let , let and fix . Let and . The LogitBoost increment of (34) is, written with ,
This is the expression the paper compares with the Gentle AdaBoost update (37).
Formalization Note The left-hand side restates the increment of (34) locally (that display belongs to the LogitBoost mission). depends on only, so ; this is part of what is to be shown.
Preamble
import Mathlib import Definitions.Def_AddLogReg_Gentle_Setting open MeasureTheory ProbabilityTheory
Formal statement
namespace AddLogReg.Gentle
theorem eq_38 {X : Type*} [MeasurableSpace X] (ν : Measure (X × Bool))
[IsProbabilityMeasure ν] (F : X → ℝ) (x : X) :
(1 / 2) * ((∫ b, (AddLogReg.LogitBoost.ystar b - AddLogReg.LogitBoost.symLogistic (F x)) ∂(ν.condKernel x)) /
∫ _b, AddLogReg.LogitBoost.symLogistic (F x) * (1 - AddLogReg.LogitBoost.symLogistic (F x)) ∂(ν.condKernel x)) =
(1 / 2) * ((AddLogReg.ExpCrit.condProb ν true x - AddLogReg.LogitBoost.symLogistic (F x)) /
(AddLogReg.LogitBoost.symLogistic (F x) * (1 - AddLogReg.LogitBoost.symLogistic (F x)))) := by sorry
end AddLogReg.Gentle
Source
Friedman, Hastie and Tibshirani, Additive logistic regression: a statistical view of boosting, Ann. Statist. 28 (2000), p. 354, §4.4, (38); p. 352, (34)
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.