ResearchTrend.AI
  • Papers
  • Communities
  • Events
  • Blog
  • Pricing
Papers
Communities
Social Events
Terms and Conditions
Pricing
Parameter LabParameter LabTwitterGitHubLinkedInBlueskyYoutube

© 2025 ResearchTrend.AI, All rights reserved.

  1. Home
  2. Papers
  3. 2011.03789
9
7

Estimation of smooth functionals in high-dimensional models: bootstrap chains and Gaussian approximation

7 November 2020
V. Koltchinskii
ArXivPDFHTML
Abstract

Let X(n)X^{(n)}X(n) be an observation sampled from a distribution Pθ(n)P_{\theta}^{(n)}Pθ(n)​ with an unknown parameter θ,\theta,θ, θ\thetaθ being a vector in a Banach space EEE (most often, a high-dimensional space of dimension ddd). We study the problem of estimation of f(θ)f(\theta)f(θ) for a functional f:E↦Rf:E\mapsto {\mathbb R}f:E↦R of some smoothness s>0s>0s>0 based on an observation X(n)∼Pθ(n).X^{(n)}\sim P_{\theta}^{(n)}.X(n)∼Pθ(n)​. Assuming that there exists an estimator θ^n=θ^n(X(n))\hat \theta_n=\hat \theta_n(X^{(n)})θ^n​=θ^n​(X(n)) of parameter θ\thetaθ such that n(θ^n−θ)\sqrt{n}(\hat \theta_n-\theta)n​(θ^n​−θ) is sufficiently close in distribution to a mean zero Gaussian random vector in E,E,E, we construct a functional g:E↦Rg:E\mapsto {\mathbb R}g:E↦R such that g(θ^n)g(\hat \theta_n)g(θ^n​) is an asymptotically normal estimator of f(θ)f(\theta)f(θ) with n\sqrt{n}n​ rate provided that s>11−αs>\frac{1}{1-\alpha}s>1−α1​ and d≤nαd\leq n^{\alpha}d≤nα for some α∈(0,1).\alpha\in (0,1).α∈(0,1). We also derive general upper bounds on Orlicz norm error rates for estimator g(θ^)g(\hat \theta)g(θ^) depending on smoothness s,s,s, dimension d,d,d, sample size nnn and the accuracy of normal approximation of n(θ^n−θ).\sqrt{n}(\hat \theta_n-\theta).n​(θ^n​−θ). In particular, this approach yields asymptotically efficient estimators in some high-dimensional exponential models.

View on arXiv
Comments on this paper