任意信源与马氏信源的比较及若干极限性质(英文)  

The comparison between an arbitrary information source and Markov chains and its limit properties

作　　者：汪忠志[1] 陈文波[1] 

机构地区：[1]安徽工业大学数理学院,安徽马鞍山243002

出　　处：《兰州大学学报（自然科学版）》2005年第2期128-133,共6页Journal of Lanzhou University(Natural Sciences)

基　　金：Supported by Anhui Higher Education Research Grant(2004Kj069)

摘　　要：引入样本散度率概念,即h(P|Q)作为任意信源与马氏信源差异的一种度量,研究了h(P|Q)与频率的强大数定理之间的联系,在适当的条件下,得到h(P|Q)的下界.We introduce the concept sample divergence rate of P relative to Q, i.e. h(P|Q) as a measure of 'dissimilarity' between arbitrary information source and Markov chain, and study the relationship between h(P|Q) and the strong law of large numbers of frequency. Under suitable restricted conditions, we get the lower bound of h(P|Q),

关 键 词：熵密度 强极限定理 样本散度率 

分 类 号：O211.4[理学—概率论与数理统计]