Optimal Estimators for Heavy-Tailed Mean Estimation via Convex Analysis
Abstract
We study optimal estimation of the location parameter of a distribution known only to lie in a symmetric moment class C0: the mean-zero distributions with bounded moment ∫ϕ\, d P B for a fixed even ϕ. Our main result concerns the fixed-margin regime, where the error margin Δ is fixed as n∞: we give an exact large-deviation characterization of the smallest worst-case probability βn(Δ) of an error exceeding Δ that any measurable estimator can guarantee with n observations. Its exponential rate is exactly a two-point Hellinger exponent over the class shifted to means Δ, r(Δ)=- PΔ∈ CΔ∫d P-Δ\, d PΔ, achieved non-asymptotically, βn(Δ) e-nr(Δ), by a monotone M-estimator synthesized from a two-parameter convex program. Lagrangian duality collapses the infinite-dimensional search over estimating functions to two multipliers, which determine a pair of envelopes characterizing the optimal estimating functions; the sandwich shape posited ad hoc in prior constructions emerges naturally. For bounded variance (ϕ(x)=x2, B=σ2) the exponent is r(Δ)=12(1+Δ2/σ2). In the fixed-confidence regime, holding β fixed and letting the optimal margin Δn(β) shrink with n, the same synthesis stays optimal to leading order for several concrete classes. As β0 it attains the sharp constant 2 of Catoni for bounded variance and the constant L(α) of Lee and Bhatt et al. for bounded α-moments, α∈(1,2), thereby shown tight; for slowly varying ϕ it is leading-order minimax at every fixed β. The least-favorable distributions are simple, supported on at most three atoms.
Turn this paper into a full lesson
ArcXiv compiles a staged curriculum from this paper: 8-12 lessons across beginner → advanced, synthesised section guides, visuals, flashcards, a quiz, exercises, and on-demand deep dives per section. Grounded in the abstract, never invented.