任意齐次树指标马氏链场的一类Shannon-Mcmillan定理

采用构造相容分布与非负上鞅的方法来研究任意齐次树指标可列非齐次马氏链场的相对熵密度的一类强极限定理,并由此得出若干齐次树指标有限状态非齐次马氏链场、一般非齐次马氏链的Shannon-Mcmillan定理.将已有的关于离散信源的结果加以推广.     

Some properties and applications of cumulative Kullback–Leibler information

The cumulative Kullback–Leibler information has been proposed recently as a suitable extension of Kullback–Leibler information to the cumulative distribution function. In this paper, we obtain various results on such a measure, with reference to its relation with other information measures and notions of reliability theory. We also provide some lower and upper bounds. A dynamic version of the cumulative Kullback–Leibler information is then proposed for past lifetimes. Furthermore, we investigate its monotonicity property, which is related to some new concepts of relative aging. Moreover, we propose an application to the failure of nanocomponents. Finally, in order to provide an application in image analysis, we introduce the empirical cumulative Kullback–Leibler information and prove an asymptotic result. Copyright © 2015 John Wiley & Sons, Ltd.     

关于样本熵的一类强偏差定理

设{Xn,n≥1}是在可数集S中取值的一列随机变量,其联合分布为gn(x1,……,xn),pk(xk)是Xk的边缘分布。本文利用关于乘积分布∏^nk=1pk(xk)的相对熵和样本相对熵率的概念建立了关于参考样本熵的一类强偏差定理。     

Re-examination of Bregman functions and new properties of their divergences

ABSTRACT

The Bregman divergence (Bregman distance, Bregman measure of distance) is a certain useful substitute for a distance, obtained from a well-chosen function (the ‘Bregman function’). Bregman functions and divergences have been extensively investigated during the last decades and have found applications in optimization, operations research, information theory, nonlinear analysis, machine learning and more. This paper re-examines various aspects related to the theory of Bregman functions and divergences. In particular, it presents many sufficient conditions which allow the construction of Bregman functions in a general setting and introduces new Bregman functions (such as a negative iterated log entropy). Moreover, it sheds new light on several known Bregman functions such as quadratic entropies, the negative Havrda-Charvát-Tsallis entropy, and the negative Boltzmann-Gibbs-Shannon entropy, and it shows that the negative Burg entropy, which is not a Bregman function according to the classical theory but nevertheless is known to have ‘Bregmanian properties’, can, by our re-examination of the theory, be considered as a Bregman function. Our analysis yields several by-products of independent interest such as the introduction of the concept of relative uniform convexity (a certain generalization of uniform convexity), new properties of uniformly and strongly convex functions, and results in Banach space theory.     

关于微分熵的一类强偏差定理

利用关于乘积分布密度的相对熵和相对熵率的概念,建立了相依连续型随机变量序列关于参考微分熵的一类强偏差定理,证明中给出了将Laplace变换应用于微分熵强偏差定理的研究的一种途径．     

任意信源关于赌博系统的一类Shannon-McMillan定理

采用网微分法和分析运算方法来研究赌博系统中任意随机变量序列随机条件熵的一类强极限定理,并由此得出若干任意信源的Shannon-Mcmillan定理.将已有的关于离散信源的结果加以推广.     

关于m阶非齐次马氏信源的一类Shannon-McMillan定理

采用构造相容分布与非负上鞅的方法的研究m阶马氏信源广义相对熵密度的强极限定理,即广义Shannon-McMillan定理.并由此得出若干马氏信源,无记忆信源的随机Shannon-Mcmillan定理.将已有的马氏信源的结果加以推广     

关于Shannon—Mcmillan定理的若干研究（Ⅱ）

该文通过概率空间上的任意分布列与独立分布列比较，研究任意随机变量序列相对熵密度用不等式给出的强极限定理，即小偏差定理，并由此得出若干Shannon－Mcmillan定理，将作者已有的关于离散信源的结果加以推广.     

A note on the existence of a largest topological factor with zero entropy

Given a topological system on a -compact Hausdorff space and its factor we show the existence of a largest topological factor containing such that for each -invariant measure , . When a relative variational principle holds, .

     

对数似然比与整值随机变量序列的一类强律

本文引进对数似然比作为整值随机变量序列相对于服从几何分布的独立随机变量序列的偏差的一种度量,并通过限制对数似然比给出了样本空间的一个子集.在此子集上得到了一类用不等式表示的强律,其中包含整值随机变量序列与相对熵密度及几何分布的熵函数有关的若干极限性质.     

关于滑动相对熵的若干极限性质

引入滑动似然比和滑动相对熵等概念,讨论了滑动相对熵的若干渐近性质.主要结果是,获得了随机变量任意状态的相对频率与滑动相对熵之间的关系,并且给出了滑动相对熵基于相对频率的上界的估计.作为推论,得到了随机序列滑动平均的强大数定理.     

关于有限马氏链相对熵密度和随机条件熵的一类极限定理

本文引进有限非齐次马链随机条件熵的概念，研究这个概念与相对熵密度的关系，并通过数列的绝对平均收敛的概念给出了有限非齐次马氏链的相对频率，相对熵密度和平均随机条件熵ａ．ｅ收敛于常数及有限非齐次马氏链熵率存在的条件。     

Relative entropy functions for factor maps between subshifts

Let and be topological dynamical systems and a factor map. A function is a compensation function for if for all . For a factor code between subshifts of finite type, we analyze the associated relative entropy function and give a necessary condition for the existence of saturated compensation functions. Necessary and sufficient conditions for a map to be a saturated compensation function will be provided.

     

关于相依离散随机序列的若干强偏差定理

首次引入随机序列滑动似然比与滑动相对熵概念作为刻画任意相依随机序列的独立逼近的随机性度量,利用B-C引理与分析方法相结合,研究任意离散随机序列部分和滑动平均的强偏差定理.作为推论,得到了关于广义经验分布函数的一个极限定理.最后,给出了若干例子.     

基于相对熵的多属性决策组合赋权方法

综合各种赋权方法给出的主观和客观属性权重信息,建立了求解多属性决策问题属性权重的优化模型,并改进了文献[11]中模型的求解方法.根据各种主客观赋权法给出的赋权结果的贴近度确定其在权重集成中的加权系数,贴近度通过计算权重向量的相对熵来得到,最后通过应用实例对此方法予以说明.     

一种基于合作与满意的群决策方法

从信息熵和相对熵的概念出发 ,提出了“不合作度”和“不满意度”的概念 ,通过论证指出了理想合作群“不合作度”和“不满意度”均为零的特点 .针对此特点 ,建立了基于“不合作度”和“不满意度”最小的优化模型 ,并针对模型 ,对遗传算法进行改进 ,使之适合于求解模型 .最后通过算例表明了此群决策方法的有效性和实用性 .     

Some characterization results on generalized cumulative residual entropy measure

The cumulative residual entropy (CRE) has been found to be a new measure of information that parallels Shannon entropy, refer to Rao et al. (2004). In this paper we study a generalized cumulative residual information measure based on Verma’s entropy function and a dynamic version of it. The exponential, Pareto and finite range distributions, which are commonly used in reliability modeling, have been characterized using this generalized measure.     

基于相对熵排序的建筑工程项目投标决策

投标决策是施工企业面对的重要问题.在建筑工程项目投标问题上,先建立风险指标体系,再结合层次分析法和信息熵法求得风险指标的综合权重,然后借助相对熵排序法,对各个方案进行综合评价,得到有效反映各方案差异的评价模型,通过对比可知此法可行有效,且灵敏度高.     

A Pathwise Approach of Some Classical Inequalities

The aim of this pedagogical paper is to show how some renowned inequalities may be obtained via a simple argument: entropy projection from the path space onto finite-dimensional coordinates spaces. Some applications are given: ergodic behaviour, perturbation.     

Some small deviation theorems for arbitrary continuous random sequence

Let (Xn)n∈N be a sequence of arbitrary continuous random variables, by the notion of relative entropy h(μ)μ(ω) as a measure of dissimilarity between probability measure μ and reference measure (μ), the explicit, general bounds for the partial sums of arbitrary continuous random variables under suitable conditions are developed. The argument uses the known and elementary lemma of convergence for likelihood ratio.