Large Deviations via Almost Sure CLT for Functionals of Markov Processes
Large Deviations via Almost Sure CLT for Functionals of Markov Processes
复制标题
通过几乎肯定的 CLT 实现马尔可夫过程泛函的大偏差
DOI：
10.1080/07362994.2012.704859
复制
发表时间：
2012
影响因子：
1.3
通讯作者：
A. Korzeniowski
中科院分区：
文献类型：
--
作者：
Adina Oprisan;A. Korzeniowski
我们考虑连续时间马尔可夫过程的加性泛函，并基于对数平均的经验测度证明了它们的泛函中心极限定理（FCLT）有几乎确定的版本。对于相应的经验过程，我们证明了基于鞅分解的大偏差原理（LDP），该原理是针对马尔可夫过程的可加泛函建立的。
We consider additive functionals of continuous time Markov processes and prove that their functional central limit theorems (FCLT) admit almost sure versions based on empirical measures with logarithmic averaging. For the corresponding empirical processes, we prove a large deviation principle (LDP) based on a martingale decomposition, established here for additive functionals of Markov processes.