Expansions in the local and the central limit theorems for dynamical systems (Q2070859)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Expansions in the local and the central limit theorems for dynamical systems |
scientific article |
Statements
Expansions in the local and the central limit theorems for dynamical systems (English)
0 references
24 January 2022
0 references
This long paper provides a detailed analysis of the asymptotic behavior of sequences of random variables, addressing the convergence issues related to the central limit theorem, its local version, and mixing properties. The main goal of the paper is to investigate the rates of convergence via expansions of arbitrary order (this refers to the existence of the finite moment of a given order), which go beyond the Berry-Esséen estimates and involves Edgeworth expansions. Some of the investigated expansions have been previously obtained for a number of chaotic non-invertible dynamical systems and bounded observables. The authors look for an extension of the results in [\textit{K. Fernando} and \textit{C. Liverani}, Ann. Inst. Henri Poincaré, Probab. Stat. 57, No. 1, 469--505 (2021; Zbl 1480.37016)] on Edgeworth expansions for weakly dependent random variables, to hyperbolic systems modeled by Young towers with exponential tails and to unbounded observables. The article is divided in two parts. In Part I, the state expansions are described in a general probabilistic context, intentionally adapted (but not restricted) to a class of dynamical systems. The existence of Edgeworth expansions is proved in that setting. In Part II, a detailed verification of probabilistic assumptions is provided in the context of hyperbolic dynamical systems. The general framework is illustrated by several examples: mixing subshifts of finite type with Lipschitz observables, systems modeled by Young towers including Sinai billiard with unbounded observables, and random matrix products. This strengthens the authors' viewpoint that in the context of dynamical systems, the study of expansions for central limit theorems and their extensions is not just a curiosity from probability theory. There are important applications. Specifically a new mechanism is identified in the decay of correlations. The mixing averaged Edgeworth expansions are deduced for a large class of nonuniformly hyperbolic dynamical systems with a large class of unbounded observables.
0 references
Birkhoff sums
0 references
Edgeworth expansions
0 references
characteristic functions
0 references
chaotic probability
0 references
Berry-Esséen estimate
0 references
central limit theorem
0 references
correlations decay
0 references
hyperbolic systems
0 references
Young towers
0 references
Sinai billiards
0 references
Hénon-type attractors
0 references
0 references
0 references
0 references
0 references
0 references
0 references