On weakly bounded empirical processes

DOI10.1007/S00208-007-0152-9MaRDI QIDQ2471759zbMATH OpenOpenAlexFDO

Authors Shahar Mendelson

Publication date 18 February 2008

Published in Mathematische Annalen (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/math/0512554

zbMATH Keywords

empirical processes

Mathematics Subject Classification ID

Geometric probability and stochastic geometry (60D05) Inequalities; stochastic orderings (60E15)

Abstract: Let $F$ be a class of functions on a probability space $(Omega,mu)$ and let $X_1,\mathrm{...},X_k$ be independent random variables distributed according to $mu$. We establish high probability tail estimates of the form $su{p}_{finF}|i:|f(X_i)|geqt$ using a natural parameter associated with $F$. We use this result to analyze weakly bounded empirical processes indexed by $F$ and processes of the form $Z_f=|k^{-1}su{m}_{i=1}^k|f{|}^p(X_i)-E|f{|}^p|$ for $p>1$. We also present some geometric applications of this approach, based on properties of the random operator $Gamma=k^{-1/2}su{m}_{i=1}^{k}inr{X}_i,cdot{e}_i$, where the $(X_i{)}_{i=1}^k$ are sampled according to an isotropic, log-concave measure on $R^n$.

Review text

Let \(X_1,\ldots,X_k\) be independent and identically distributed random variables with sample space \(\Omega\), and let \(F\) be a class of functions defined on \(\Omega\). The main result is an upper bound on \(\sup_{f\in F}| \{i:| f(X_i)| \geq t\}| \) for every \(t>0\) which holds with probability close to one. The bound depends on certain geometric characteristics of \(F\). The result is used to analyze \(\sup_{f\in F}\big| k^{-1}\sum_{i=1}^k| f| ^p(X_i)-E| f| ^p\big| \) for \(p>1\) in terms of the geometry of \(F\). Geometric applications are given when \(X_1,\dots,X_k\) are sampled from an isotropic, log-concave measure on \(R^n\).

Recommendations

• Empirical processes with a bounded \(\psi_1\) diameter
• Sharper bounds for Gaussian and empirical processes
• scientific article; zbMATH DE number 51427
• Inégalités exponentielles pour les processus empiriques
• Bounding the expectation of the supremum of an empirical process over a (weak) VC-major class

Cites work

• \(L_{p}\)-moments of random vectors via majorizing measures
• A probabilistic approach to the geometry of the \(\ell^n_p\)-ball
• Concentration of mass on convex bodies
• Concentration property on probability spaces.
• Global versus local asymptotic theories of finite-dimensional normed spaces
• scientific article; zbMATH DE number 3931824 (Why is no real title available?)
• scientific article; zbMATH DE number 49190 (Why is no real title available?)
• scientific article; zbMATH DE number 194093 (Why is no real title available?)
• scientific article; zbMATH DE number 1302647 (Why is no real title available?)
• Majorizing measures: The generic chaining
• Non commutative Khintchine and Paley inequalities
• Orlicz norms of sequences of random variables
• Random vectors in the isotropic position
• Random walks and anO*(n5) volume algorithm for convex bodies
• Reconstruction and subgaussian processes
• Regularity of Gaussian processes
• Sharper bounds for Gaussian and empirical processes
• Some limit theorems for empirical processes (with discussion)
• Subspaces of Small Codimension of Finite-Dimensional Banach Spaces
• The concentration of measure phenomenon
• The Generic Chaining
• Weak convergence and empirical processes. With applications to statistics

Cited in

(16)

• Sharper bounds for Gaussian and empirical processes
• On least squares estimation under heteroscedastic and heavy-tailed errors
• Bounding the expectation of the supremum of an empirical process over a (weak) VC-major class
• Upper bounds on product and multiplier empirical processes
• Quantitative estimates of the convergence of the empirical covariance matrix in log-concave ensembles
• On multiplier processes under weak moment assumptions
• Empirical processes with a bounded \(\psi_1\) diameter
• Dimension-free bounds for sums of independent matrices and simple tensors via the variational principle
• Matrix deviation inequality for ℓp-norm
• A Bernstein-type inequality for suprema of random processes with applications to model selection in non-Gaussian regression
• Sharp concentration of simple random tensors
• Extreme eigenvalues of log-concave ensemble
• Weak limits of empirical measures for a family of diffusion processes
• Approximating the moments of marginals of high-dimensional distributions
• Aggregation via empirical risk minimization
• Regularization in kernel learning

This page was built for publication: On weakly bounded empirical processes

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2471759)