CASIA OpenIR  > 毕业生  > 博士学位论文
Alternative TitleThe Analysis and Application of the Information Integration Mechanism in Social Tagging Systems
Thesis Advisor戴汝为
Degree Grantor中国科学院研究生院
Place of Conferral中国科学院自动化研究所
Degree Discipline模式识别与智能系统
Keyword社会化标注系统 非负矩阵分解 标签语义挖掘 资源分类显示与浏览 Social Tagging System Non-negative Matrix Factorization Tag Semantics Resource Classifying And Exploring
Abstract随着万维网的发展,以Delicious、Flickr、豆瓣网、、citeulike等为代表的社会化标注系统日益流行,其集资源的存储与分享、资源的社会化标注与评论、用户社区等多种功能于一体,允许用户将感兴趣的资源(网页、照片、图书、电影、音乐、学术内容等)收藏到网站中,并为其自由添加关键词作为标签。用户与系统的交互过程中形成了大量用户产生的数据,这些数据显示了用户对于标注资源的喜好,体现了用户对标注资源语义信息的高度概括,带有用户的主观认知特点,反映了用户的相关知识结构,大量用户的标注行为呈现出社会化协作的特性。因此,如何充分利用这些标注数据为用户提供准确的信息服务是学术界及企业界关注的热点。社会化标注系统包含资源、用户、标签、评论与评价等元素,如何集成化考虑这些元素,深入理解社会化标注系统;如何解决由于用户的自由标注导致的标签的冗余性和语义模糊性问题;如何帮助用户更方便地浏览及检索资源,都是当前亟待解决的课题。围绕这些问题,本文的主要工作与贡献有如下几个方面: 1. 从整体性、复杂性的角度深入研究社会化标注系统的信息集成机制。首先,应用马斯洛需求层次理论描述标注系统中的用户需求,接着用标注系统中多个空间概括资源、用户、标签以及近年来涌现出的新元素,进而探讨标注系统通过资源、用户、标签、评价与评论、知识等多个空间的集成与交互,帮助用户群体实现由数据、信息到知识以及智慧的认知过程。在此基础上,围绕更好地满足用户需求,提出使用户与系统紧密结合、以用户为中心的系统设计与计算的理念。 2. 针对标签的自由与非控制性导致其在使用上存在的冗余和语义模糊性问题,提出一种基于非负矩阵分解(Non-negative Matrix Factorization)的语义相关标签挖掘方法,针对用户收藏的资源与用户使用的标签间的关系,定义了资源与标签间的关系集合,借助非负矩阵分解算法从标签空间中发现标签间的潜在语义相关性。 3. 针对用户浏览收藏的网络资源时遇到的问题,提出一个基于用户标注的信息分类显示与浏览系统框架,通过挖掘用户标签空间中语义相关的标签集合,帮助用户自动构建语义相关的“标签包”,并利用互联网开放的第三方资源组成的知识库,为“标签包”赋予相应的类别名称,方便用户按照“类别名”分类浏览资源及相关的标签集合。同时,还考虑用户保存资源的时间信息,挖掘出用户收藏兴趣的变迁。 基于以上研究,研制了一个资源和标签的分类显示与浏览原型系统,为提高标注系统服务质量,满足用户需求,进行了初步的探索和实践。
Other AbstractWith the rapid development of the World Wide Web, social tagging systems are becoming more and more popular, such as Delicious, Flickr, Douban,, citeulike, and so on, which provide functions including storage, sharing, social tagging, reviewing, rating, and constructing user communities. Tagging systems allow users to save resources (i.e. web pages, photos, books, movies, music, science-related content, etc.) into web sites, and choose keywords as tags freely. During the interaction process between users and systems, a large number of user-generated data are available, which show users’ interests and represent their summarization for the semantic information in the resources. The user behavior data reflect their subjective cognitive characteristics and relevant knowledge structure, and present social collaborative characteristics. Therefore, how to utilize such data to provide services for users becomes the focus of attention by the academia and business world. Social tagging systems involve many elements such as resources, users, tags, reviews, ratings, and so on. How to integrate all these elements to deepen the understanding of social tagging systems, how to help users deal with the tag redundancy and semantic ambiguity problem, and how to efficiently browse and search their resources are essential tasks. The main work and contributions of the thesis are summarized as follows. 1. The information integration mechanism of social tagging systems is studied from a holism and complexity perspective. The needs of users are described by applying the Maslow’s hierarchy of needs theory, then multiple spaces are used to reflect elements in social tagging systems, such as resources, users, tags, reviews, rating, knowledge, and so on. The purpose is to help users achieve the cognitive emergence process from data, information to knowledge and intelligence through the integration and interaction among different spaces in tagging systems. Based on such analysis, we propose the idea of a user-system-cooperated user-centered systematic design and its computation for better satisfying the needs of users. 2. To deal with the problems of tag redundancy and semantic ambiguity , a semantic related tag mining approach based on non-negative matrix factorization method is proposed. We define relational sets between resources and tags to discover the latent semantic relevance among tags in the tag space from a perspective of the relationship between users' resour...
Other Identifier200718014628075
Document Type学位论文
Recommended Citation
GB/T 7714
张雷鸣. 社会化标注系统中信息集成机制分析及其应用[D]. 中国科学院自动化研究所. 中国科学院研究生院,2010.
Files in This Item:
File Name/Size DocType Version Access License
CASIA_20071801462807(8373KB) 暂不开放CC BY-NC-SAApplication Full Text
Related Services
Recommend this item
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[张雷鸣]'s Articles
Baidu academic
Similar articles in Baidu academic
[张雷鸣]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[张雷鸣]'s Articles
Terms of Use
No data!
Social Bookmark/Share
All comments (0)
No comment.

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.