An application of the information theory to filtering problems

The purpose of this paper is to study the filtering problems from the viewpoint of the information theory. For a linear system it is proved that the necessary and sufficient condition for maximizing the mutual information between a state and the estimate is to minimize the entropy of the estimation error. Then we derive the Kalman-Bucy filter for both the discrete-time and the continuous-time systems by an application of the information theory. Furthermore, the approach is extended to the nonlinear dynamical systems with noisy observations and then the information structures of the optimal filter for a continuous-time nonlinear system are made clear, which has been presented as the interesting open problems by Bucy.     

改进的模糊遗传算法及在信息过滤中的应用

为了改进传统遗传算法在求解复杂问题上存在早收敛及搜索后期运行效率低等缺点,提出了一种应用于文本分类和信息过滤的模糊遗传算法.首先应用了年龄概念来控制种群规模,使得遗传操作过程更接近于自然进化过程,然后引进参数的模糊调整过程,对遗传算法的参数种群规模,交叉率及变异率3个方面进行动态调整,改进了遗传算法的搜索性能.实验结果表明,相比传统遗传算法,该模糊遗传算法在全局优化能力及收敛速度上均有显著提高.     

模糊信息粗传递模型及其应用

在实际的信息传递中人们需要更加准确可靠的信息,而且被传递概念是模糊集合,根据这些信息传递情况,利用粗模糊集理论提出了一种模糊信息粗传递模型,由该模型得到提高模糊信息传递精确度方法。最后给出了该模型的应用。模糊粗传递是粗模糊集的一种新的应用。     

网络信息重构模型的优选算法及其改进

从信息资源角度出发,针对网络信息发布过程的问题进行研究,建立信息重构模型,并根据简化模型进行优选方案讨论。从时间复杂度考虑,使用动态规划原理深入研究,对优选算法进行改进,得到网络信息重构模型优选算法的改进方案,使该算法得到更完善的发展,在网络信息重构发布方面拥有更广泛的实用性。     

A metagenetic algorithm for information filtering and collection from the World Wide Web

This paper describes the implementation of evolutionary techniques for information filtering and collection from the World Wide Web. We consider the problem of building intelligent agents to facilitate a person's search for information on the Web. An intelligent agent has been developed that uses a metagenetic algorithm in order to collect and recommend Web pages that will be interesting to the user. The user's feedback on the agent's recommendations drives the learning process to adapt the user's profile with his/her interests. The software agent utilizes the metagenetic algorithm to explore the search space of user interests. Experimental results are presented in order to demonstrate the suitability of the metagenetic algorithm's approach on the Web.     

基于退火遗传算法的网络信息过滤系统研究

分析了网络信息过滤一般模型以及现有技术,研究了如何更准确地构建用户模板,提出了一种基于遗传算法的网络信息过滤系统模型,并且引入了遗传扩展操作和Boltzmann群体更新准则来改进遗传算法存在的缺点,同时给出了一种Roocchio反馈模型对用户兴趣模板进行更新和维护.实验结果表明,基于该模型设计的网络信息过滤系统能够有效实现对网络信息过滤.     

A flexible multi criteria information filtering model

In this paper, a flexible multi criteria information filtering model is presented. This model is flexible since it allows choosing several distinct criteria, such as content aboutness, coverage, novelty, trust, timeliness, and combining them by a soft aggregation to define a personalized filter. The personal filter is encoded into the user profile that also contains the representations of the user interests that can evolve over time. An implementation of the system applying a combination of the aboutness and coverage criteria has been evaluated and compared to other filtering systems, showing its superior effectiveness. Finally, the possible use of the other criterion is discussed.     

A two-stage decision model for information filtering

Information mismatch and overload are two fundamental issues influencing the effectiveness of information filtering systems. Even though both term-based and pattern-based approaches have been proposed to address the issues, neither of these approaches alone can provide a satisfactory decision for determining the relevant information. This paper presents a novel two-stage decision model for solving the issues. The first stage is a novel rough analysis model to address the overload problem. The second stage is a pattern taxonomy mining model to address the mismatch problem. The experimental results on RCV1 and TREC filtering topics show that the proposed model significantly outperforms the state-of-the-art filtering systems.     

Approximate nonlinear filtering and its application in navigation

In this paper, we introduce for the first time particle filtering for an exponential family of densities. We prove that under certain conditions the approximated conditional density of the state converges to the true conditional density. In the realistic setting where the conditional density does not lie in an exponential family but stays close to it, we show that under certain assumptions the error of the estimate given by an approximate nonlinear filter (which we call the projection particle filter), is bounded. We use projection particle filtering in state estimation for a combination of inertial navigation system (INS) and global positioning system (GPS), referred to as integrated INS/GPS. We illustrate via numerical experiments that projection particle filtering outperforms regular particle filtering in navigation performance, and extended Kalman filter as well when satellite loss-of-lock occurs.     

Organizing domain-specific information on the Web: An experiment on the Spanish business Web directory

Web directories organize voluminous information into hierarchical structures, helping users to quickly locate relevant information and to support decision-making. The development of existing ontologies and Web directories either relies on expert participation that may not be available or uses automatic approaches that lack precision. As more users access the Web in their native languages, better approaches to organizing and developing non-English Web directories are needed. In this paper, we have proposed a semi-automatic framework, which consists of anchor directory boosting, meta-searching, and heuristic filtering, to construct domain-specific Web directories. Using the framework, we have built a Web directory in the Spanish business (SBiz) domain. Experimental results show that the SBiz Web directory achieved significantly better recall, F-value, efficiency, and satisfaction rating than the benchmark directory. Subjects provided favorable comments on the SBiz Web directory. This research thus contributes to developing a useful framework for organizing domain-specific information on the Web and to providing empirical findings and useful insights for end-users, system developers, and researchers of Web information seeking and knowledge management.     

An application of the cusp catastrophe model to user information satisfaction

This study shows that there is a nonlinear relationship between overall user information satisfaction (UIS) and an index created by combining satisfaction responses to various information systems components/attributes. Several instruments have been developed in the IS literature to measure UIS. These measure UIS by asking respondents to rate their satisfaction/dissatisfaction with a variety of IS attributes; for example, EDP services, EDP staff, information products, and involvement in IS development. Tests of predictive validity, in the past, have shown only a 55% correlation between global UIS and UIS scores. Here, we show that the global UIS measure and IS factor scores follow a nonlinear relationship and are, in fact, cusp-distributed. The extent of use of IS is shown to be a splitting factor in this distribution and produces abrupt, catastrophic changes. Responses from faculty members at an academic institution were used to test for the distribution using IS component scores as the normal independent variable and the extent of use of IS as the splitting variable. The cusp model was a more efficient predictor than the linear model or a control equation using bimodal transformation.     

基于自助采样和互信息策略的集成校正算法及应用

综合应用训练集自助采样(bootstrap)和互信息(mutual information)选择变量来引入成员模型间的差异性,提出一种子空间回归的集成校正算法ESPLS.当建立一成员模型时,先淘汰互信息量小于一个特定阈值的变量,使建模在原变量的一个子空间上进行,有效避免了多元共线性产生的诸多问题.通过一近红外光谱数据集实验,同时与全谱偏最小二乘法(PLS)和互信息选择变量的偏最小二乘法(SPLS)2种单模型算法进行了比较,证明:该算法在不增加模型复杂度的前提下,能提高校正模型的预测精度、稳定性及抗过拟合的能力.     

Web社会网络的粗糙属性图模型及应用

针对Web环境下的社会网络具有信息粗糙性的特征,即Web数据中有大量垃圾内容和垃圾链接,同时很多信息是不完整的、缺失的,且信息有重复现象存在等,在已提出的属性图模型基础上,结合粗糙集理论解决不完备信息的优势,首先提出粗糙顶点属性图和粗糙边属性图,进而给出粗糙属性图的概念,以对Web社会网络结构进行分析,使其能够描述复杂Web社会网络中的不完整信息以及动态变化的链接。其次对粗糙属性图的粗糙特性进行分析,给出粗糙顶点精度、粗糙边精度和粗糙图精度等概念,得出粗糙属性图的精度与顶点和边集属性划分程度有关的结论,即人们对图的认知程度与图的精度密切相关。最后,在中国知网上通过对论文作者进行查询得到粗糙图,并通过不断添加顶点属性,将图顶点划分得越来越精细,挖掘出要查询的作者合作关系图,从而说明粗糙属性图在社会网络分析中符合人们的认知过程。     

粒子滤波理论及其在目标跟踪中的应用

非线性估计领域的经典算法是扩展Kalman滤波（EKF）,它采用了Taylor展开的线性变换来近似非线性模型,因而存在计算量大、实时性差、估计精度低等缺点。而粒子滤波采用一些带有权值的随机样本（粒子）来表示所需要的后验概率密度,而不是采用传统的线性变换,从而得到基于物理模型的近似最优数值解,具有精度高、收敛速度快等特点。对经典的纯方位跟踪问题进行了仿真。仿真结果表明,粒子滤波器的跟踪性能要远优于EKF的性能。     

一种基于事件的Web程序测试模型

基于状态转换的测试方法是探测Web程序动态行为的有效途径。针对Web应用中复杂的动态页面,通过改进现有Web测试方法,提出了一种新的基于事件的测试模型,用Web关系图（WARD）描述Web系统的结构,用基于事件的状态迁移图（SMD）描述复杂动态页面的内部结构,两者结合对Web应用系统全面建模。在实际Web系统中的实验证明该方法的建模是全面准确的。     

基于规则模型的网页主题文本提取方法

通过对网页结构化和半结构化信息的分析,提出了一种基于规则模型的网页正文提取方法.该方法在总结HTML标签的不同应用特征和网页布局的结构特征的基础上,通过定义一系列过滤、提取和合并规则来建立一个通用的网页正文抽取模型,以达到有效提取网页主题文本的目的.实验结果表明,该方法对于各类型网页主题文本的提取均具有较高的准确卒,通用性强.     

Rama: An architecture for Internet information filtering

This paper describes Rama, a first generation experimental information retrieval and filtering system that attempts to recover useful information from various Internet sources including USENIX news and anonymous FTP servers. The focus of the Rama system to date has been on building a distributed query and information retrieval system, which provides an interface to heterogeneous information services. A user of Rama sends one or more asynchronous queries to a Rama server using existing SMTP e-mail clients. The server periodically searches local and remote Internet services. Searches are prefiltered with the use of timestamps. Data objects which are newer than the timestamp are then searched via a query mechanism which relies on a combination of vector-distance, pattern matching operands, and boolean operators. Results are weighted according to how closely they match queries and are posted via e-mail to the user. Input to the e-mail client can be further filtered — one can use the MH mail system and sort input by weight. Results indicate that the current system is useful and extensible. So far we have assumed that existing e-mail systems will be used for input and output and have not attempted to construct special client interfaces. Efforts are underway to extend the system with WWW searching capabilities and construct a special WWW oriented user-interface.     

Huber-Based滤波及其在相对导航问题上的应用

研究了一种考虑视觉导航设备输出噪声不符合理想高斯分布时,无人机编队INS/Vision相对导航方法。给出了相对惯导方程的推导过程和相对视线矢量测量原理,应用Huber-Based滤波融合相对惯导信息和相对视线矢量信息,估计出无人机之间的相对姿态、相对位置和相对速度。Huber-Based滤波是一种混合 , 范数最小估计,对受到污染的高斯分布噪声,该估计方法具有一定的鲁棒性。最后,通过仿真研究验证了相对导航方法的有效性。     

基于Web挖掘的个性化算法及其在网络教学平台的应用

在Web挖掘的基础上设计针对Web服务的Web访问事务模型WTM和个性化推荐算法。算法以WTM为基础,旨在根据用户的访问模式向用户推荐个性化的Web资源。其利用关联规则得到的频繁项集实时地匹配用户的当前访问序列,对不同的用户提供不同的推荐资源。在此过程中不需产生所有的关联规则,提高了推荐的效率。最后,将该模型和算法应用于网络教学实践得出了个性化的网络教学环境。     

高斯混合交互式多模型容积信息滤波算法

针对机动目标跟踪中由于目标机动使系统的非线性强度增大,导致系统的线性误差增大和跟踪精度明显下降、甚至发散的问题,提出了基于高斯混合的交互式多模型容积信息滤波（ GMIMM-CIF）算法,实现对机动目标的精确跟踪。新算法在每次输入交互之后,保留概率较大的几个假设,并利用一个高斯混合项替换最优多模型算法中剩余的假设,从而使算法中假设的数量保持恒定;用容积信息滤波器（ CIF）代替传统的非线性滤波器,通过估计信息状态向量和信息矩阵而不是估计状态向量和协方差,可以减小系统的非线性误差。通过仿真对比实验,验证了该算法可以提高机动目标的跟踪精度。