Suppose that X₁,…,Xₙ are independent, identically distributed random variables of mean zero and variance one. Assume that E\, |X₁|⁴ ≤ δ⁴. We observe that there exist many choices of coefficients θ₁,…,θₙ ∈ R with ∑ⱼ θⱼ² = 1 for which $_{α, β ∈ { R} {α < β}} | { P} ( α ≤ ∑ⱼ₌₁^n θ_j X_j ≤ β ) -({√2 π})-1/2 ∫αβ e-t^2/2\,dt | ≤ {C δ^4}/{n},$ where $C> 0$ is a universal constant. This inequality should be compared with the classical Berry--Esseen theorem, according to which the left-hand side may decay with $n$ at the slower rate of $O( 1/ √n)$ for the unit vector $θ = (1,…,1) / √n$. An explicit, universal example for coefficients $θ = (θ_1,…,θ_n)$ for which this inequality holds is $θ = {(1,√2, -1, -√2, 1, √2, -1, -√2, …)}({3n/2})-1/2,$ when $n$ is divisible by four. Parts of the argument are applicable also in the more general case, in which $X_1,…,X_n$ are independent random variables of mean zero and variance one yet are not necessarily identically distributed. In this general setting, the bound above holds with $δ^4 = n⁻¹ ∑ⱼ₌₁^n { E}\, |X_j|^4$ for most selections of a unit vector $θ = (θ_1,…,θ_n) ∈ { R}^n$. Here “most” refers to the uniform probability measure on the unit sphere.
No takes yet. Share an insight, caveat, or question.
Klartag et al. (2011) studied this question.
Synapse has enriched 4 related papers. Consider them for comparative context: