提取文本流主题的神经网络新算法

目前，关于动态文本数据处理已逐渐成为数据挖掘的研究热点，例如，在聊天室中提取热门主题以及所有的讨论主题。目前已有的神经网络方法能较好地提取所讨论的主题，但不能决定哪个主题是热门主题，而且，提取到的主题之间相互干扰。利用主题之间相互独立和主题自相关的特性，基于自相关矩阵以及独立主元分析教学模型，本文提出一种新的神经网络方法，该算法能成功解决这些问题。在Yahoo聊天室上的实验结果表明，本文算法能准确提取主题以及热门主题，并且主题之间相互干扰大大减小。关键词独立主元分析，神经网络，自相关矩阵，时间序列     

基于视觉特征的BBS用户发言信息抽取

针对传统从BBS主题网页中自动抽取用户发言信息的问题,提出独立于网页脚本语言的自动抽取方法,可根据网页提供的视觉信息进行有效抽取.通过归纳BBS主题网页内部用户发言信息的视觉特征,并以此为依据指导整个抽取过程.实验结果表明,基于视觉的抽取方法可以得到较高的抽取质量.     

基于复杂网络社区结构的论坛热点主题发现

社区结构是复杂网络的重要特征之一,该文通过构造基于兴趣的论坛用户网络,成功地将社区结构发现的理论与方法应用于论坛热点主题的自动发现,提出了极大社区的概念和反复挖掘极大社区的方法,发现网络结构和论坛主题在反复挖掘过程中均几乎保持不变。仿真结果表明,提取效果令人满意。     

基于Web的群体研讨系统

基于Web的群体研讨系统,不仅支持成员跨平台研讨,而且能够对研讨过程中产生的发言信息进行有效管理,并支持共识达成。本文以综合集成方法论为指导,给出基于Web的群体研讨平台的理论基础、系统建模、系统框架及系统实现方法。以“我国新农合发展所存在的问题”为主题,在建成的平台上进行模拟研讨,并对结果进行分析。一方面验证了系统的可行性,另一方面为新农合相关主题的进一步深入研究奠定理论基础。     

基于微粒群算法的组织评估优化方法及实例

社区结构是复杂网络的重要特征之一,该文通过构造基于兴趣的论坛用户网络,成功地将社区结构发现的理论与方法应用于论坛热点主题的自动发现,提出了极大社区的概念和反复挖掘极大社区的方法,发现网络结构和论坛主题在反复挖掘过程中均几乎保持不变.仿真结果表明,提取效果令人满意.     

面向综合集成研讨环境的主动信息获取方法*

针对在综合集成研讨环境中,由于存在时间压力,传统的网络信息获取方法难以直接使用,提出了一种面向综合集成研讨环境的主动信息获取方法。该方法将领域词条与通用词条相结合,从发言文本流中实时提取话题,并在话题发生变化时自动生成检索词送入搜索引擎进行检索,再通过多个用户之间的协作推荐实现对重要检索结果的筛选。实验结果表明,这种方法可为综合集成研讨系统的用户提供及时、准确、上下文相关的信息服务。     

企业竞争情报主题挖掘与主题演化研究

研究了我国企业竞争情报的热点主题和主题演化态势,利用主题挖掘与主题演化方法系统梳理了我国企业竞争情报领域的研究成果.通过Python自动提取及预处理文献数据,再利用共词分析、LDA模型和知识图谱挖掘该领域的核心科研群体和热点主题,最后结合主题演化方法梳理企业竞争情报的发展脉络.该研究可为企业竞争情报领域今后的相关探索提供借鉴,具有一定的应用价值.     

利用组合模型生成微博热点话题事件摘要

针对微博热点话题检测使用主题模型只能提取出无序话题词组合的问题,提出一种结合词激活力模型与主题模型各自优点的微博热点话题检测方法及话题关键词的计算方法。首先,使用传统的主题模型提取出微博文本中的热点主题;其次,根据各主题下文档的概率分布提取出新的话题文档;然后引入词激活力模型计算各个词之间的词激活力,生成词激活力矩阵;最后,利用词激活力矩阵生成有序的词序列作为热点事件。实验验证了该方法的可行性,表明所提出方法能够很好地识别出热点词并生成可读性高的事件。     

一种基于即时发言评价的专家权威度计算方法

针对综合集成研讨环境中缺少具体的专家群体交互和权威度计算方法这一问题,提出一种基于即时发言评价的专家权威度计算方法.通过计算专家在研讨过程中的发言质量和评价质量,得到专家的总体权威度.根据专家个体之间的发言评价响应关系建立专家群体交互结构.实验表明该方法计算结果科学合理,为与会专家在综合集成研讨环境中进行交互提供良好条件.     

文本主题的自动提取方法研究与实现

在深入分析了当前流行的文本主题提取技术和方法的基础上，将语义方法融入统计算法，提出了一种基于统计的主题提取方法，并描述了它的实现过程。该方法利用文档内句子之间的语义相关性，实现了文本主题的自动生成。首先对文本进行切词和分句处理实现信息分割，再结合文本聚类技术对文本句进行聚类实现信息合并，最后从每类中抽取代表句生成文本主题。实验结果表明，该方法是一个有效、实用的方法。     

The Hot List Strategy

Experimentation strongly suggests that, for attacking deep questions and hard problems with the assistance of an automated reasoning program, the more effective paradigms rely on the retention of deduced information. A significant obstacle ordinarily presented by such a paradigm is the deduction and retention of one or more needed conclusions whose complexity sharply delays their consideration. To mitigate the severity of the cited obstacle, I formulated and feature in this article the hot list strategy. The hot list strategy asks the researcher to choose, usually from among the input statements characterizing the problem under study, one or more statements that are conjectured to play a key role for assignment completion. The chosen statements – conjectured to merit revisiting, again and again – are placed in an input list of statements, called the hot list. When an automated reasoning program has decided to retain a new conclusion C – before any other statement is chosen to initiate conclusion drawing – the presence of a nonempty hot list (with an appropriate assignment of the input parameter known as heat) causes each inference rule in use to be applied to C together with the appropriate number of members of the hot list. Members of the hot list are used to complete applications of inference rules and not to initiate applications. The use of the hot list strategy thus enables an automated reasoning program to briefly consider a newly retained conclusion whose complexity would otherwise prevent its use for perhaps many CPU-hours. To give evidence of the value of the strategy, I focus on four contexts: (1) dramatically reducing the CPU time required to reach a desired goal, (2) finding a proof of a theorem that had previously resisted all but the more inventive automated attempts, (3) discovering a proof that is more elegant than previously known, and (4) answering a question that had steadfastly eluded researchers relying on an automated reasoning program. I also discuss a related strategy, the dynamic hot list strategy (formulated by my colleague W. McCune), that enables the program during a run to augment the contents of the hot list. In the Appendix, I give useful input files and interesting proofs. Because of frequent requests to do so, I include challenge problems to consider, commentary on my approach to experimentation and research, and suggestions to guide one in the use of McCunes automated reasoning program OTTER.     

常用循环摘要的自动生成方法及其应用

采用形式化方法证明软件的正确性是保障软件可靠性的有效方法,而对循环语句的分析与验证是形式化证明中的关键,对循环语句的处理一直是程序分析与验证中的一个难点问题.本文提出使用循环语句修改的内存和这些内存中存放的新值来描述循环语句的执行效果,并将该执行效果定义为循环摘要.同时,本文提出了一种自动生成循环摘要的方法,可以为操作常用数据结构的循环自动生成循环摘要,包含嵌套循环.此外,基于循环摘要,我们可以自动生成循环语句的规约,包括循环不变式、循环的前置条件以及循环的后置条件.我们已经实现了自动生成循环摘要以及循环规约的方法,并将它们集成到验证工具Accumulator中,实验表明,我们的方法可以有效地生成循环摘要,并生成多种类型的规约,从而辅助软件程序的形式化证明,提高验证的自动化程度和效率,减轻验证人员的负担.     

Auto-encoder based bagging architecture for sentiment analysis

Sentiment analysis has long been a hot topic for understanding users statements online. Previously many machine learning approaches for sentiment analysis such as simple feature-oriented SVM or more complicated probabilistic models have been proposed. Though they have demonstrated capability in polarity detection, there exist one challenge called the curse of dimensionality due to the high dimensional nature of text-based documents. In this research, inspired by the dimensionality reduction and feature extraction capability of auto-encoders, an auto-encoder-based bagging prediction architecture (AEBPA) is proposed. The experimental study on commonly used datasets has shown its potential. It is believed that this method can offer the researchers in the community further insight into bagging oriented solution for sentimental analysis.     

消息传递并行程序的弱变异测试及其转化

并行程序执行的不确定性,增加了测试的复杂性和难度.研究消息传递并行程序的变异测试,提出其弱变异测试转化方法,以提高该程序变异测试的效率.首先,根据消息传递并行程序包含语句的类型和语句变异之后导致的变化构建相应的变异条件语句;然后,将构建好的所有变异条件语句插入到原程序中,形成新的被测程序,从而将原程序的弱变异测试问题转化为新程序的分支覆盖问题.这样做的好处是,能够利用已有的分支覆盖方法解决变异测试问题.将该方法应用于8个典型的消息传递并行程序测试中,实验结果表明,该方法不但是可行的,也是必要的.     

Using Metaphor to Analyse Qualitative Data: Vulcans and Humans in Software Development

This paper reports on an experience of using metaphor in qualitative research of software engineering in practice. Our project aimed to uncover non-technical factors affecting the adoption and evolution of Software Quality Management Systems (referred to here as ‘the quality process’). Previously we have reported the tensions we uncovered around the quality process in four companies, based on semi-structured interviews. This paper extends this work by applying metaphor to the results. We show how we were able to produce more general statements regarding the tensions and their amelioration, and then introduce results from a fifth company, which we compare against our general statements. We find that these statements are generally supported by results from this fifth company. Finally we present some reflections on our experience of using metaphor in this way.     

The power of combining resonance with heat

In this article, I present experimental evidence of the value of combining two strategies each of which has proved powerful in various contexts. The resonance strategy gives preference (for directing a program's reasoning) to equations or formulas that have the same shape (ignoring variables) as one of the patterns supplied by the researcher to be used as a resonator. The hot list strategy rearranges the order in which conclusions are drawn, the rearranging caused by immediately visiting and, depending on the value of the heat parameter, even immediately revisiting a set of input statements chosen by the researcher; the chosen statements are used to complete applications of inference rules rather than to initiate them. Combining these two strategies often enables an automated reasoning program to attack deep questions and hard problems with far more effectiveness than using either alone. The use of this combination in the context of cursory proof checking produced most unexpected and satisfying results, as I show here. I present the material (including commentary) in the spirit of excerpts from an experimenter's notebook, thus meeting the frequent request to illustrate how a researcher can make wise choices from among the numerous options offered by McCune's automated reasoning program OTTER. I include challenges and topics for research and, to aid the researcher, in the Appendix a sample input file and a number of intriguing proofs.This work was supported by the Mathematical, Information, and Computational Sciences Division subprogram of the Office of Computational and Technology Research, U.S. Department of Energy, under Contract W-31-109-Eng-38.     

Unimodular transformations of non-perfectly nested loops

A framework is described in which a class of imperfectly nested loops can be restructured using unimodular transformations. In this framework, an imperfect loop nest is converted to a perfect loop nest using Abu-Sufah's Non-Basic-to-Basic-Loop transformation. Conditions for the legality of this transformation and techniques for their verification are discussed. An iteration space, which extends the usual concept so as to represent explicitly the executions of individual statements, is proposed to model the converted loop nest. Since the converted loop nest is a perfect loop nest, data dependences can be extracted and optimal transformations can be selected for parallelism and/or locality in the normal manner. To generate the restructured code for a unimodular transformation, a code generation method is provided that produces the restructured code that is free of if statements by construction.     

基于Android平台的通讯录中文汉字的排序研究

随着移动互联网的迅速发展,Android平台也得到迅速普及,该移动平台上的应用日臻完善。通讯录作为人们之间联系方式存放的介质,自然成为必不可少的应用。而其中联系人的显示顺序则成为了开发者备受关注的话题,提出了一种新的排序方法,借助数据库的排序语句,方便高效地实现了将汉字按照拼音首字母自增的方式进行排序,该方法使用起来便捷、高效,避免了程序中引入第三方jar包所导致的程序臃肿问题。     

Efficient Index Set Generation for Compiling HPF Array Statements on Distributed-Memory Machines

In languages such as High Performance Fortran (HPF), array statements are used to express data parallelism. In compiling array statements for distributed-memory machines, efficient enumeration of local index sets and commmunication sets is important. A method based on a virtual processor approach has been proposed for efficient index set enumeration for array statements involving arrays distributed using block-cyclic distributions. The virtual processor approach is based on viewing a block-cyclic distribution as a block (or cyclic) distribution on a set of virtual processors, which are cyclically (or block-wise) mapped to the physical processors. The key idea of the method is to first develop closed forms in terms of simple regular sections for the index sets for arrays distributed using block or cyclic distributions. These closed forms are then used with the virtual processor approach to give an efficient solution for arrays with the block-cyclic distribution. HPF supports a two-level mapping of arrays to processors. Arrays are first aligned with a template at an offset and a stride and the template is then distributed among the processors using a regular data distribution. The introduction of a nonunit stride in the alignment creates “holes” in the distributed arrays which leads to memory wastage. In this paper, using simple mathematical properties of regular sections, we extend the virtual processor approach to address the memory allocation and index set enumeration problems for array statements involving arrays mapped using the two-level mapping. We develop a methodology for translating the closed forms for block and cyclically distributed arrays mapped using a one-level mapping to closed forms for arrays mapped using the two-level mapping. Using these closed forms, the virtual processor approach is extended to handle array statements involving arrays mapped using two-level mappings. Performance results on the Cray T3D are presented to demonstrate the efficacy of the extensions and identify various trade-offs associated with the proposed method.