【作者】 周姗姗；

【导师】 毕强；

【作者基本信息】 吉林大学 ， 情报学， 2014， 博士

【摘要】 随着网络技术的应用与普及，数字资源的开发利用已成为时代赋予的重大课题。众所周知，数字资源组织是数字资源开发利用的前提，然而事实上数字资源组织往往呈现出局部有序而整体无序的窘境，具体体现为数字资源孤岛、数字资源超载等，这大大阻碍了用户对数字资源的有效获取和共建共享。为摆脱这种窘境，学者们都将解决该问题的途径聚焦在了数字资源的再组织上。关于数字资源的再组织，国内学者在长期的探索实践中依次形成了“资源整合”和“资源聚合”两种模式。数字资源聚合尚未形成完备的方法体系，但也探寻出了两类行之有效的操作行方法：一为知识组织方法；二为计量分析方法。知识组织方法以概念分析、本体、关联数据等方法增强资源语义，计量分析方法以共现关系、耦合关系、社会网络分析等方法挖掘资源关联。目前，数字资源聚合的途径主要通过结构维度下的领域本体来实现，其语义化程度较高但却无法摆脱概念树结构的桎梏。而关联维度则能够突破主题领域内部严谨苛刻的层级限制，基于知识间的关联关系将平面树形结构拓展到网络立体空间。群聚维度更是将知识的客观属性与人类的主观认知相结合，充分体现了知识与人的交互。笔者认为，以多维度融合的视角考查数字资源聚合，有助于突破知识组织研究领域的瓶颈，使基于复杂网络分析和Folksonomy（分众分类法）融合的数字资源多维度聚合成为可能。鉴于此，本研究尝试将分众分类法和复杂网络分析二者有机地融合在一套体系之下，提出基于复杂网络分析的分众分类法模式下数字资源多维度聚合的思想，探索提高现代知识服务的解决方案。本研究遵循“理论—模型—实证”的技术路线，研究聚焦于复杂网络与分众分类法互补融合视角下的数字资源“多维度聚合”，从结构维度、群聚维度、关联维度三个方面论述了数字资源的多维性，构建了数字资源聚合的多维度理论框架，主要研究内容包括如下：（1）通过对与选题相关的国内外文献的分析与综述，确定了本论文研究的逻辑主线：数字资源聚合的基本逻辑——结构维度；数字资源聚合的社会逻辑——群聚维度；数字资源聚合的扩展逻辑——关联维度。（2）从资源聚合的理念、维度、实现技术和具体应用等四个方面对分众分类法和复杂网络分析进行比较分析，力图寻找分众分类法与复杂网络分析在数字资源聚合方向上的融合点。（3）通过对用户标签网络的属性分析以及模块化处理，将标签云中的标签划分成若干个知识群落。借助连线、颜色、字号的相互配合，从主题知识关联的视角优化标签云，为标签云体系研发提供依据。（4）运用复杂网络分析方法对基于网络中心性的标签关系、基于网络群聚性的标签关系、基于网络关联性的标签关系，以及基于复杂网络的分众分类法扩展等问题进行系统研究。（5）从知识的结构维度、知识的群聚维度、知识的关联维度等方面对基于复杂网络分析的分众分类法数字资源聚合进行实证研究。更多还原

【Abstract】 With the popularization and application of the network technology, the developmentand utilization of digital information resources have become a major task endowed by thetimes. As everyone knows, the organization of digital resource is the prerequisite for thedevelopment and utilization of digital resources, but in fact the organization of digitalresource often presents the partial ordered and disordered situation overall, embodied in thedigital resource, digital resource overload, which greatly hinders the acquisition and sharingof the digital resources of users. In order to get rid of the dilemma, all the scholars havefocused on how to reorganize digital resources to solve this problem. In the long-termexploration and practice, domestic scholars have in turn formed two such modes on how toreorganize the digital resources as the "integration of resources" and "resource sharing". Acomplete system of polymerization method of digital resources has not been formed, butthey have also explored two kinds of effective methods of operation: knowledgeorganization method and quantitative analysis method. The former one mainly uses conceptanalysis, ontology, relational data and other methods to enhance the resource semantics; thelatter one mainly uses co-occurrence, coupling relationship, social network analysis methodof mining resources association.At present, the digital resources aggregation approach and knowledge discovery isrealized by the dimensions of the organizational structure of the domain ontology. Thesemantic degree is high but can’t get rid of the shackles of concept tree structure. Thecorrelation dimension is able to break through the rigorous and harsh topic hierarchy limit.The two-dimensional tree structure is extended to three-dimensional network space basedon the relationship between knowledge. Transfer diffusion dimension is the subjectivecognition and objective attribute of human knowledge combination, which fully embodiesthe interaction between knowledge and human. The author believes that, in order toexamine the digital resources integration of multidimensional perspective of polymerization,which helps to break through the bottleneck of the field of knowledge organization andresearch, the complex network analysis and Folksonomy fusion based on degree of polymerization is possible. In view of this, this study attempts to organically combineFolksonomy and complex network analysis into a set of system, puts forward the idea ofmultidimensional digital resources of degree of polymerization of the complex networkanalysis based on Folksonomy mode, and explores the solution to improve the modernknowledge service.This study follows the technical route of the "theory model empirical", focuses on thecomplex network and Folksonomy complementary digital resources" under the perspectiveof the integration of the multidimensional polymerization", discusses the multidimensionaldigital resources from the dimensions of organization structure, spreads three aspectsdimension and correlation dimension and constructs a theoretical framework ofmultidimensional digital resources polymerization. The main research contents are asfollows:(1) Through the analysis and summary of the topics related with the domestic andforeign literature, the logical clue of this thesis is the basic logic structure dimension ofdigital resources polymerization; digital resources aggregation of social logic clusterdimension; the extended logic digital resources aggregation associated dimension.(2) From the four aspects of resource aggregation concept, dimension, implementationtechnology and application this study conducts comparative analysis of Folksonomy andcomplex network analysis and tries to find the Folksonomy and complex network analysispolymerization fusion point in the direction of digital resources.(3) Through the attribute analysis of the user label network and modularized, the tagcloud label is divided into a number of knowledge communities. By the cooperation withline, color, font size, from the perspective of optimizing the tag cloud theme knowledgerelated, it provides the basis for the tag cloud system R&D.(4) By using complex network analysis method, this study conducts research ondimensional expanding of Folksonomy based on network centrality of labels, networkcluster of labels, and network connection of labels,.(5) From the three dimensions of structure dimension, cluster dimension and associated dimension, this study conducts research on digital resources aggregation based onFolksonomy of the complex network.更多还原