首页 | 本学科首页   官方微博 | 高级检索  
相似文献
 共查询到20条相似文献,搜索用时 269 毫秒
1.
提出了将支持向量机应用于文本自动分类的研究,与常用的K-最邻近法相比,无论是对训练数据集还是测试数据集均具有一定的优势,而且不同特征选择方法对支持向量机的影响要比K-最邻近法小。此外,从研究中的不同特征选择的评价函数来看,它们对分类有一定的影响,应用X2统计进行特征选择的分类正确率最高,其次是文本证据权,而期望交叉熵的效果最差,说明特征选择在文本自动分类中也是相当重要的。  相似文献   

2.
Web自动文本分类技术研究综述   总被引:1,自引:0,他引:1  
Web自动文本分类是信息检索与数据挖掘领域的研究热点与核心技术,近年来得到了广泛的关注和快速的发展.本文首先分析了国内外Web自动文本分类方法的研究现状,接着对新近出现的多分类器融合的方法、基于群的分类方法、基于RBF网络的文本分类模型、基于模糊-粗糙集的文本分类模型、潜在语义分类模型等新方法,以及K-近邻算法和支持向量机的新发展等进行了深入探讨;并对Web自动文本分类过程中的几个关键技术:文本预处理、文本表示、特征降维、训练方法和分类算法等进行了分析;最后总结了当前Web自动文本分类技术存在的问题及其发展趋势.  相似文献   

3.
将自动文本分类引入竞争情报系统,并结合民航客服业的实际应用场景展开研究。在对文本分类的理论基础进行研究与分析的基础上,设计一种自动分类策略,其核心是针对传统机器学习方法依赖历史数据的局限,优化特征选择和学习样本获取方法。选用SVM算法,详细描述分类词表构建、分类规则提取、分类模型训练等过程,最终取得较为理想的实际效果。  相似文献   

4.
王煜  白石  王正欧 《情报学报》2007,26(5):643-647
本文提出了一种基于权重优化的样本相似度测量的距离公式,改进了KNN文本分类算法.KNN算法通常采用传统的VSM模型,各个特征具有相同的权重,使其不适应于文本处理的环境.本文首先根据神经网络理论,采用灵敏度方法对文本特征向量的每个特征的权重进行修正,并且采用降低运算量的神经网络特征选择方法进行第二次降维处理.然后根据同一特征对不同类别的文本类的分类作用不同,对距离公式中的特征权重进行进一步改进,从而进一步提高了KNN文本分类算法的精度.  相似文献   

5.
赵洪 《情报学报》2020,(3):330-344
自动文摘是文本挖掘的主要任务之一。相比于抽取式自动文摘,生成式自动文摘在思想上更接近人工摘要的过程,具有重要研究意义。近几年伴随着深度学习方法的发展,基于深层神经网络模型的生成式自动文摘也有了令人瞩目的发展。为了更全面地理解该类方法的思想和研究现状,本文从生成式自动文摘的任务描述入手,梳理了基于RNN (recurrent neural network,循环神经网络)的模型、基于CNN (convolutional neural network,卷积神经网络)的模型、基于RNN+CNN的模型、融合注意力机制的模型和融合强化学习的模型共五大类生成式自动文摘的深度学习方法。这类方法表明,在深层神经网络的训练下,特别是融合注意力机制和强化学习后,摘要效果得以明显提升。在生成式自动文摘研究的未来发展中,除深度学习方法本身的不断应用和改进外,还需关注如何有效实现篇章级语义理解下的摘要、面向不同文本对象特点的摘要和摘要结果自动评价等问题。此外,如何结合传统摘要研究中的成熟方法进一步提高摘要效果,也是一个很有价值的研究方向。  相似文献   

6.
将神经网络集成思想引入WEB文本分类领域,构造一个用于Web文本分类的多BP神经网络集成模型;详述模型的设计思路与结构框架,并分别在公有的英文数据集、实际的中文数据集上进行分类实验;与经典的SVM模型、KNN模型相比,神经网络集成模型具有更高的分类精度,且对于训练样本集规模具有更好的鲁棒性,不失为一种高效的文本分类新方法,研究其在文本分类领域的应用将是一个有前景的方向。  相似文献   

7.
基于词典约简及多分类算法的文本分类系统的设计与开发   总被引:3,自引:1,他引:2  
文本自动分类是目前机器学习、自然语言处理和信息资源检索领域的研究热点之一.本文在对自动分类的实现技术问题进行探讨的基础上,尝试对自动分类的几个环节提出改进措施.具体包括:采用约简法进行抽词词典的构造,采用投票法进行文本特征的选择,采用层次法进行逐层次的分类,采用统计与规则相结合的方法进行分类器的构造等.通过在不同语料库上与传统分类方法的对比测试表明,上述改进措施能够有效提高自动分类的性能,基于这些改进措施所开发的自动分类系统具有大规模文本自动分类的可行性.本文详细讨论了相关改进措施的原理、实现算法、流程及存在的问题.  相似文献   

8.
根据互信息、RBF神经网络和关联规则原理,提出了一种抽取WEB文本分类规则的新方法。先根据互信息选择和各类相关程度大的若干词条,然后采用RBF神经网络方法对选择的特征进行进一步提取,得到维数较小的文本特征向量空间。之后再根据挖掘出的关联规则获取WEB文本分类规则,建立文本分类器,在保证了分类精度的前提下抽取出利于理解的文本分类规则。  相似文献   

9.
Web内容挖掘在数字图书馆中的应用   总被引:3,自引:1,他引:3  
黎琳  赵英 《图书馆学研究》2006,(2):19-21,24
Web信息是数字图书馆数字资源的重要来源,文章详细阐述了对Web内容中的文本信息的挖掘,包括:文本自动摘要、文本分类和文本聚类。在挖掘过程中基于用户需求和用户特征,针对文本分类,重点分析了分类过程和在数字图书馆中的应用;针对文本聚类,介绍两种基本聚类原理以及与文本分类的不同之处,重点论述其在数字图书馆中如何应用。最后提出内容挖掘与用户挖掘的结合更有助于服务用户。  相似文献   

10.
微博短文本预处理及学习研究综述   总被引:1,自引:0,他引:1  
王连喜 《图书情报工作》2013,57(11):125-131
认为因短文本具有特征稀疏性和高度冗余性,微博短文本的预处理及学习方法研究已经成为微博信息挖掘及应用的关键,并在许多方面有着非常重要和广泛的应用。重点分析微博短文本的特性,并对微博短文本的预处理和学习方法及其应用现状进行归纳和总结,包括短文本特征表示、短文本特征拓展与选择、短文本分类与聚类学习、热点事件发现及自动文摘等。最后指出相关研究的局限性,并对未来的发展方向进行展望。  相似文献   

11.
This article surveys a sample of sources of the information about Romania available to British readers in nineteenth century British newspapers and periodicals. It traces first contacts between the Romanian lands and Britain after the union of the principalities of Wallachia and Moldavia in 1859, then after their independence from the Ottoman Empire. The article highlights an increased Romanian interest in British periodicals, which reported and reviewed Romanian literature and scholarship. The article concludes that nineteenth century British newspapers and periodicals offer a great variety and wealth of new material previously unavailable or unknown to researchers. It also states that only a portion of a large quantity of this material has been indexed and is therefore available via the bibliographic sources mentioned in the article. The author argues for the need of a new and updated British-Romanian bibliography, which can draw on new online resources offering access to thousands of new newspapers and periodical records.  相似文献   

12.
ABSTRACT

The paper looks at library approval plans for material published in Slavic, East European, and Eurasian countries from the selector's point of view. Reasons why a selector would or would not want one are examined. Success with approval plans requires monitoring receipts, as well as good and ongoing communication among the selector, the acquistions department, and the vendor. A preliminary list of vendors offering approval plans for the countries of the region appears in the appendix.  相似文献   

13.
为进一步提升武汉科技信息共享服务平台使用效率,本文从平台资源建设、资源应用、供需对接方式和供需特点等方面分析了武汉科技信息资源服务现状;基于需求和利用的角度,结合平台管理实践和走访用户、问卷调查等研究方法,从信息资源需求主体和平台自身建设管理两个维度,找出制约科技资源供需对接的主要因素;以市场化和制度化为创新理念,从政策创新、机制创新、市场化服务、环境营造、人才培养等方面提出平台建设由“资源集聚”向“需求导向”转变的对策建议.  相似文献   

14.
ABSTRACT

The article examines the most important periodicals of ethnic minorities in Poland. After 1989, many ethnic groups (e.g., Germans and Romanies) were allowed to publish journals and newspapers for the first time since the end of World War II. The publications examined show the rich cultural life of the various ethnic groups as well as their current status in Poland. In addition to popular titles, some scholarly publications are also discussed.  相似文献   

15.
The author answers a reference question on bibliographic sources for the Ukrainian periodical press 1840–1850. Helpful publications include bibliographies, guides, and library catalogs. These potentially make mention of revolutionary developments in Hungary (such as the Twelve Points paragraph of the Demands of the Hungarian Nation in March 1848, the subsequent April Laws, and Hungary's declaration of independence in April 1949), and elsewhere in the Hapsburg Empire.  相似文献   

16.
在分析文献在不同研究阶段用词时间特征的倾向性基础上,提出一种基于主题模型的研究发展阶段识别方法。重点阐述该方法的构建过程,包括时间特征抽取、发展阶段界定、主题冷热变化分析等步骤。为验证该方法的有效性,针对词频统计法和主题模型方法在主题演化分析中的效果进行比较分析。结果表明,该方法能在识别主题热点和发展趋势的同时,有效地区分不同主题所反映的研究发展阶段。  相似文献   

17.
Chromebooks and the G Suite group of products, like Google Search, Gmail, and Google Docs, have rapidly expanded in American schools during the past 5 years. The impact of one-to-one Chromebook devices and the pervasive use of Google's software products in American education cannot be overstated. This article explores some of the influences of these products on research, based on the experiences of a librarian and technology coordinator at an elementary education level. The author has several suggestions for effective research with these products in mind.  相似文献   

18.
Library historians can learn a great deal from studying the gender of authorship and institutional affiliation of a scholarly journal. The focus of this study is to examine these two aspects of journal production in library history to see who is producing published research in this field. Twenty-three years of Libraries & Culture were chosen as the target volumes. The study reveals that more men than women published in library history as well as locating which institutions were represented. This type of information is usedful to the library historian engaged in the analysis of published scholarship, and more generally to scholars with an interest in patterns of literature production in fields closely related to the social and behavioral sciences.  相似文献   

19.
ABSTRACT

The history of the almanac in Croatia is reconstructed through primary research in bibliographic and archival sources. The almanac is a vehicle for knowledge communication in informal contexts, engaging both oral tradition and literary forms traceable to medieval literacy and ways of structuring knowledge. The history of the almanac in Croatia reflects the changing context of the book trade, literacy, and the evolution of language. Four main stages are identified: (1) the beginning of the annual almanac in the seventeenth century; astrological almanacs reflecting the sensibility of the Baroque period; (2) the Enlightenment's stimulation of almanac publishing in the spirit of contemporary secular reforms in agriculture and education; (3) nineteenth-and twentieth-century almanac trade, showing complex and overlapping networks for the production, distribution and appropriation of printed almanacs;(4) roughly the end of World War II, when the almanac slowly moved out of the role of a popular mass medium and into specialized niches represented by regional, diaspora, and religious almanacs.  相似文献   

20.
《The Reference Librarian》2013,54(74):121-164
Summary

Although the Internet provides access to a wealth of information, there is little, if any, control over the quality of that information. Side-by-side with reliable information, one finds disinformation, misinformation, and hoaxes. The authors of this paper discuss numerous examples of fabricated historical information on the Internet (ranging from denials of the Holocaust to personal vendettas), offer suggestions on how to evaluate websites, and argue that these fabrications can be incorporated into bibliographic instruction classes.  相似文献   

设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号