Some dichotomy results for functionals of Harris recurrent Markov chains
Let \((E,\mathfrak E)\) be a measurable space with a countably generated \(\sigma \)-algebra \(\mathfrak E\), \(\{X_{n}\}\) a Harris recurrent Markov chain in \((E,\mathfrak E)\) with a transition probability \(P\) and an invariant measure \(\pi \). Let \(B\) be a separable Banach space, \(\xi : E\rightarrow B\) a measurable map, set \(S_k = \sum ^{k}_{j=1} \xi (X_{j})\), \(k\geq 1\). In the study of limit behaviour of the sequence \(\{S_{k}\}\), various integrability properties and tail estimates appear, involving a set \(A\) from a suitable class (small sets, \(D\)-sets), cf. the author's book ``Limit theorems for functionals of ergodic Markov chains with general state space (1999; Zbl 0952.60014). In the paper, it is shown that several conditions of this type are satisfied either for all sets in the considered class, or for none. To be more explicit, let us quote one of the obtained results. Set \(\mathfrak E^{+} = \{A\in \mathfrak E; \pi (A)>0\}\), \(\tau _{A} = \inf \{n\geq 1; X_{n}\in A\}\) for any \(A\in \mathfrak E^{+}\), denote by \(\mathfrak S\) the class of all finite unions of small sets (\(S\in \mathfrak E^{+}\) is called a small set, if \(P^{m}\geq \beta \mathbf 1_{S}\otimes \nu \) for some \(m\geq 1\), \(\beta >0\) and a probability measure \(\nu \)), and by \(\widetilde {\mathfrak S}\) the subset of all \(A\in \mathfrak S\) such that \(\int _{A} E_{x}\varphi ( |S_{\tau _{A}} |) \text d\pi (x) < \infty \), where \(\varphi : [0,\infty [\to [0,\infty [\) is a nondecreasing function, \(\varphi (cs) \leq c^\lambda \varphi (s)\) for some \(\lambda >0\) and all \(s\geq 0\), \(c\geq 2\). If \(\int _{E} \varphi ( |\xi (x) |) \text d\pi (x) <\infty \), then either \(\widetilde {\mathfrak S} =\emptyset \), or \(\widetilde {\mathfrak S} = \mathfrak S\). Moreover, a sufficient condition (uniform ergodicity) for the latter case to occur is given.
- A uniform theory for sums of Markov chain transition probabilities
- Extrapolation and interpolation of quasi-linear operators on martingales
- General Irreducible Markov Chains and Non-Negative Operators
- scientific article; zbMATH DE number 3762951 (Why is no real title available?)
- scientific article; zbMATH DE number 3462924 (Why is no real title available?)
- scientific article; zbMATH DE number 3236503 (Why is no real title available?)
- Large deviations for vector-valued functionals of a Markov chain: Lower bounds
- Limit theorems for functionals of ergodic Markov chains with general state space
- Loi fonctionnelle du logarithme itéré pour les processus de Markov récurrents. (Functional iterated logarithm law for recurrent Markov processes)
- Markov additive processes. I: Eigenvalue properties and limit theorems
- Markov additive processes. II: Large deviations
- Markov chains and stochastic stability
- Moderate deviations for empirical measures of Markov chains: Lower bounds
- Moderate deviations for empirical measures of Markov chains: Upper bounds
- On additive functionals of Markov chains
- Uniform lower bounds for randomly stopped Banach space valued random sums
- A new approach to the study of Harris type Markov operators
- Chung's law for additive functionals of positive recurrent Markov chains
- A regularity condition and a limit theorem for Harris ergodic Markov chains
- scientific article; zbMATH DE number 5018838 (Why is no real title available?)
- Limit Theorems for Harris Markov Chains, II
This page was built for publication: Some dichotomy results for functionals of Harris recurrent Markov chains
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1613637)
