Processes with long memory: Regenerative construction and perfect simulation (Q1872393)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Processes with long memory: Regenerative construction and perfect simulation |
scientific article |
Statements
Processes with long memory: Regenerative construction and perfect simulation (English)
0 references
6 May 2003
0 references
The authors are concerned with at most countable state space infinite order chains, a special case of random systems with complete connections; see, e.g., Section 5.5 in [\textit{M. Iosifescu} and \textit{S. Grigorescu}, ``Dependence with complete connections and its applications'' (1990; Zbl 0749.60067)]. Their main result is an explicit construction, as a deterministic function of a sequence of i.i.d. random variables uniformly distributed over the unit interval, of a stationary process which is an infinite order chain. As corollaries they present (1) an alternative proof of the existence and uniqueness of the stationary process, and (2) a perfect simulation algorithm and a regeneration scheme for this process. [The first regenerative structures for infinite order chains were proposed by S. P. Lalley (in 1986 and 2000) and \textit{H. Berbee} (in 1997) under assumptions which are stronger than those of the present authors.] The authors' approach leans heavily on that of \textit{P. A. Ferrari}, \textit{A. Maas}, \textit{S. MartÃnez} and \textit{P. Ney} [Ergodic Theory Dyn. Syst. 20, No. 6, 1657-1670 (2000; Zbl 0992.37011)], which in turn shares results with \textit{H. Berbee} [Probab. Theory Relat. Fields 76, 243-253 (1987; Zbl 0611.60059)] and \textit{P. Ney} and \textit{E. Nummelin} [ibid. 96, No. 4, 503-520 (1993; Zbl 0792.60078)]. As applications, the authors discuss the perfect simulation of Markov chains on the unit interval and of binary autoregressive processes.
0 references
simulation algorithm
0 references
regeneration scheme
0 references
perfect simulation of Markov chains
0 references
binary autoregressive processes
0 references