Self-normalized processes: exponential inequalities, moment bounds and iterated logarithm laws.
From MaRDI portal
Publication:1879810
DOI10.1214/009117904000000397zbMATH Open1075.60014arXivmath/0410102OpenAlexW3101257511MaRDI QIDQ1879810FDOQ1879810
Authors: Victor H. de la Peña, Michael J. Klass, Tze Leung Lai 
Publication date: 15 September 2004
Published in: The Annals of Probability (Search for Journal in Brave)
Abstract: Self-normalized processes arise naturally in statistical applications. Being unit free, they are not affected by scale changes. Moreover, self-normalization often eliminates or weakens moment assumptions. In this paper we present several exponential and moment inequalities, particularly those related to laws of the iterated logarithm, for self-normalized random variables including martingales. Tail probability bounds are also derived. For random variables B_t>0 and A_t, let Y_t(lambda)=exp{lambda A_t-lambda ^2B_t^2/2}. We develop inequalities for the moments of A_t/B_{t} or sup_{tgeq 0}A_t/{B_t(log log B_{t})^{1/2}} and variants thereof, when EY_t(lambda )leq 1 or when Y_t(lambda) is a supermartingale, for all lambda belonging to some interval. Our results are valid for a wide class of random processes including continuous martingales with A_t=M_t and B_t=sqrt < M>_t, and sums of conditionally symmetric variables d_i with A_t=sum_{i=1}^td_i and B_t=sqrtsum_{i=1}^td_i^2. A sharp maximal inequality for conditionally symmetric random variables and for continuous local martingales with values in R^m, mge 1, is also established. Another development in this paper is a bounded law of the iterated logarithm for general adapted sequences that are centered at certain truncated conditional expectations and self-normalized by the square root of the sum of squares. The key ingredient in this development is a new exponential supermartingale involving sum_{i=1}^td_i and sum_{i=1}^td_i^2.
Full work available at URL: https://arxiv.org/abs/math/0410102
Recommendations
Inequalities; stochastic orderings (60E15) Martingales with discrete parameter (60G42) Martingales with continuous parameter (60G44)
Cites Work
- A general class of exponential inequalities for martingales and ratios
- Title not available (Why is that?)
- Upper bounds for the \(L_ p\)-norms of martingales
- A comparison theorem on moment inequalities between negatively associated and independent random variables
- Title not available (Why is that?)
- Self-normalized laws of the iterated logarithm
- When is the Student \(t\)-statistic asymptotically standard normal?
- The Berry-Esseen bound for Student's statistic
- Title not available (Why is that?)
- Some extensions of the LIL via self-normalizations
- Boundary Crossing Probabilities for the Wiener Process and Sample Sums
- Self-normalized Cramér-type large deviations for independent random variables.
- A Delicate Law of the Iterated Logarithm for Non-Decreasing Stable Processes
- On the LIL for self-normalized sums of IID random variables
- Estimation of densities and applications
- Self-normalized large deviations
- Maximal inequalities and the law of the iterated logarithm
- Another note on the Borel-Cantelli lemma and the strong law, with the Poisson approximation as a by-product
- Title not available (Why is that?)
- Probabilistic behavior of harmonic functions
- Inequalities for a Pair of Processes Stopped at a Random Time
- Title not available (Why is that?)
- Boundary crossing probabilities for sample sums and confidence sequences
- Title not available (Why is that?)
- Improved ratio inequalities for martingales
Cited In (30)
- Exponential inequalities for self-normalized martingales with applications
- Matrices -- compensating the loss of anschauung
- Spiking the random matrix hard edge
- Time-uniform, nonparametric, nonasymptotic confidence sequences
- Long time dynamics for interacting oscillators on graphs
- Exponential inequalities and the law of the iterated logarithm in the unbounded forecasting game
- Precise asymptotics in complete moment convergence for self-normalized sums
- Solving optimal stopping problems via empirical dual optimization
- Erdős-Feller-Kolmogorov-Petrowsky law of the iterated logarithm for self-normalized martingales: A game-theoretic approach
- An exponential inequality for autoregressive processes in adaptive tracking
- Adaptive deep learning for nonlinear time series models
- Precise asymptotics in the deviation probability series of self-normalized sums
- Time-uniform Chernoff bounds via nonnegative supermartingales
- Berry-Esseen bounds for self-normalized sums of locally dependent random variables
- Exact inequalities for sums of asymmetric random variables, with applications
- Sequential testing for elicitable functionals via supermartingales
- A law of large numbers for interacting diffusions via a mild formulation
- Theory and applications of multivariate self-normalized processes
- Encounters with Martingales in Statistics and Stochastic Optimization
- Title not available (Why is that?)
- Semi-parametric contextual bandits with graph-Laplacian regularization
- Exponential inequalities for unbounded functions of geometrically ergodic Markov chains: applications to quantitative error bounds for regenerative Metropolis algorithms
- An almost sure central limit theorem for self-normalized weighted sums
- Woodroofe's one-armed bandit problem revisited
- On the bias, risk, and consistency of sample means in multi-armed bandits
- Drift estimation for a multi-dimensional diffusion process using deep neural networks
- Tightened exponential bounds for discrete-time conditionally symmetric martingales with bounded jumps
- Invariance principles for adaptive self-normalized partial sums processes.
- On the convex Poincaré inequality and weak transportation inequalities
- Cramér-type large deviations for samples from a finite population
This page was built for publication: Self-normalized processes: exponential inequalities, moment bounds and iterated logarithm laws.
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1879810)
