Gaussian concentration bounds for stochastic chains of unbounded memory
From MaRDI portal
Abstract: We obtain optimal Gaussian concentration bounds (GCBs) for stochastic chains of unbounded memory (SCUMs) on countable alphabets. These stochastic processes are also known as "chains with complete connections" or "-measures". We consider two different conditions on the kernel: (1) when the sum of its oscillations is less than one, or (2) when the sum of its variations is finite, i.e., belongs to . We also obtain explicit constants as functions of the parameters of the model. The proof is based on maximal coupling. Our conditions are optimal in the sense that we exhibit examples of SCUMs that do not have GCB and for which the sum of oscillations is strictly larger than one, or the variation belongs to for any . These examples are based on the existence of phase transitions. We also extend the validity of GCB to a class of functions which can depend on infinitely many coordinates. We illustrate our results by three applications. First, we derive a Dvoretzky-Kiefer-Wolfowitz type inequality which gives a uniform control on the fluctuations of the empirical measure. Second, in the finite-alphabet case, we obtain an upper bound on the -distance between two stationary SCUMs and, as a by-product, we obtain new (explicit) bounds on the speed of Markovian approximation in . Third, we obtain exponential rate of convergence for Birkhoff sums of a certain class of observables.
Recommendations
- Uniform Chernoff and Dvoretzky-Kiefer-Wolfowitz-type inequalities for Markov chains and related processes
- Concentration of measure inequalities for Markov chains and \(\Phi\)-mixing processes.
- Concentration of Markov chains with bounded moments
- Subgaussian concentration inequalities for geometrically ergodic Markov chains
- Concentration bounds for entropy estimation of one-dimensional Gibbs measures
Cites work
- A note on the Bramson-Kalikow process
- A quantitative McDiarmid's inequality for geometrically ergodic Markov chains
- An example of non-unique g -measures
- Attractive regular stochastic chains: perfect simulation and phase transition
- Bounding \(\bar d\)-distance by informational divergence: A method to prove measure concentration
- Chains with complete connections and one-dimensional {G}ibbs measures
- Chains with complete connections: general theory, uniqueness, loss of memory and mixing properties
- Chains with unbounded variable length memory: perfect simulation and a visible regeneration scheme
- Concentration bounds for entropy estimation of one-dimensional Gibbs measures
- Concentration inequalities for dependent random variables via the martingale method
- Concentration inequalities for functions of Gibbs fields with application to diffraction and random Gibbs measures
- Concentration inequalities for Markov chains by Marton couplings and spectral methods
- Concentration inequalities for random fields via coupling
- Concentration inequalities. A nonasymptotic theory of independence
- Concentration of measure inequalities for Markov chains and \(\Phi\)-mixing processes.
- Convergence properties of functional estimates for discrete distributions
- Coupling and perturbation techniques for categorical time series
- Criteria for $\bar {d}$-continuity
- Decay of correlations for non Hölderian dynamics. A coupling approach
- Dynamic uniqueness for stochastic chains with unbounded memory
- Entropy Bounds for Discrete Random Variables via Maximal Coupling
- Explicit estimates in the Bramson-Kalikow model
- Exponential inequalities and estimation of conditional probabilities
- Exponential integrability and transportation cost related to logarithmic Sobolev inequalities
- Fluctuations of the Empirical Entropies of a Chain of Infinite Order
- Gaussian concentration and uniqueness of equilibrium states in lattice systems
- Gibbs measures and phase transitions.
- scientific article; zbMATH DE number 4170917 (Why is no real title available?)
- scientific article; zbMATH DE number 3945130 (Why is no real title available?)
- scientific article; zbMATH DE number 918233 (Why is no real title available?)
- Large deviations for empirical entropies ofg-measures
- Markov approximation of chains of infinite order in the \(\bar d\)-metric
- Markov approximations of chains of infinite order
- Markov Chains
- Markov chains and stochastic stability
- Measure concentration for a class of random processes
- Non-uniqueness for specifications in \(\ell^{2+\epsilon}\)
- Nonuniqueness in \(g\)-functions
- On categorical time series models with covariates
- On chains of infinite order
- On non-regular g-measures
- Perfect simulation for locally continuous chains of infinite order
- Principe variationnel et syst�mes dynamiques symboliques
- Random Markov processes and uniform martingales
- Ruelle's Operator Theorem and g-Measures
- Speed of \(\overline d\)-convergence for Markov approximations of chains with complete connections. A coupling approach
- Square summability of variations of g-functions and uniqueness of g-measures
- Strongly mixing g-measures
- Subgaussian concentration inequalities for geometrically ergodic Markov chains
- Sur des chaînes à liaisons complètes
- The positive-divergence and blowing-up properties
- Uniform Chernoff and Dvoretzky-Kiefer-Wolfowitz-type inequalities for Markov chains and related processes
- Uniqueness vs. non-uniqueness for complete connections with modified majority rules
This page was built for publication: Gaussian concentration bounds for stochastic chains of unbounded memory
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6187463)
