Stochastic Processes and their Applications ( IF 1.4 ) Pub Date : 2021-04-28 , DOI: 10.1016/j.spa.2021.04.009 Cheng-Der Fuh
Let be a complete separable metric space and a sequence of random functions from to . Motivated by studying the stability property for Markovian dynamic models, in this paper, we assume that the random function is driven by a Markov chain . Under some regularity conditions on the driving Markov chain and the mean contraction assumption, we show that the forward iterations , , converge weakly to a unique stationary distribution for each , where denotes composition of two maps. The associated backward iterations are almost surely convergent to a random variable which does not depend on and has distribution . Moreover, under suitable moment conditions, we provide estimates and rate of convergence for and , . The results are applied to the examples that have been discussed in the literature, including random coefficient autoregression models and recurrent neural network.
中文翻译:
马尔可夫迭代函数系统的渐近行为
让 是一个完全可分离的度量空间, 来自的随机函数序列 至 。通过研究马尔可夫动力学模型的稳定性,本文假设随机函数 由马尔可夫链驱动 。在驱动马尔可夫链的某些规律性条件下和平均收缩假设下,我们证明了正向迭代, ,微弱地收敛到唯一的平稳分布 对于每个 , 在哪里 表示两个地图的组成。相关的向后迭代 几乎可以肯定地收敛于随机变量 不依赖于 并有分布 。此外,在合适的时刻条件下,我们提供了以下估计和收敛速度: 和 , 。将结果应用于文献中已讨论的示例,包括随机系数自回归模型和递归神经网络。