【作者】 李敏；

【导师】 白华；

【作者基本信息】 郑州大学 ， 图书馆学， 2012， 硕士

【摘要】 一直以来,图书馆和情报领域的专家学者,都在关注各类检索语言共同应用于同一检索系统或不同检索系统,而导致检索策略多样化,进而影响到系统内信息资源共享的问题,并积极探讨解决方法。检索语言标准化应是最理想的方法,但难度太大。通过人工或自动方式,在异构检索语言词汇之间建立语义映射关系,以此实现它们的兼容转换,最终让各种检索语言集成,这是较为现实和可行的办法。在我国,《中国图书馆分类法》是影响最大、用户群最多的综合性图书分类法,如何与中外其他检索语言兼容受到了更多关注,只有融入到检索语言这个大系统中去,《中国图书馆分类法》才能继续保持其在知识组织系统中的优势地位。本文为此分五部分展开了比较系统的研究。第一部分,概述检索语言兼容的相关基础问题。首先是概念的辨析,在追溯检索语言兼容、兼容性、兼容化、兼容转换、互操作、映射及映射转换等概念的基础上,厘清它们之间的关系,进而归纳整理出狭义和广义的检索语言兼容概念。其次,从检索语言基本原理的一致性和检索语言构成方法上的互补性两个角度论述了检索语言兼容的理论基础。最后分析了影响检索语言兼容的因素：检索语言编制原则,类表词汇(类目概念)主题范畴、专指度、先组度,类表结构相似程度等。第二部分,《中图法》兼容化的历程。具体研究了《中图法》编制之初对国内外分类法的借鉴和吸收、基于手工研制的《中图法》的系列化类表和分类主题一体化词表、基于计算机技术的《中图法》的互操作研制、《中图法》与其他知识组织系统的集成。第三部分,《中图法》兼容的范围与模式。在全面调查和分析《中图法》与其他检索语言兼容的成果、研究项目和计划的基础上,指出《中图法》目前成型的兼容成果主要集中在与其系列化类表、与《汉语主题词表》兼容,同时在《中国分类主题词表》的基础上也开始了与自然语言的兼容研究,与国外分类语言、主题语言的兼容大多局限于部分领域或主题,并还需进一步论证。针对以上兼容范围总结了《中图法》在实现兼容化的过程中所采用的各种兼容模式的特点和技术性能。第四部分,网络环境下《中图法》兼容化发展中存在的问题及对策。随着国外的一些著名分类语言兼容范围扩大,检索语言兼容化发展趋势愈加突出：组织对象扩大化和进一步细化,组织工具的用途多元化,概念体系范围扩大化,结合或强化传统的知识分类和词汇控制的手段,知识组织工具多元化存在,越来越明显跨学科发展特征等。《中图法》要成为汉语界通用性的检索语言并与国外其他检索语言兼容,还面临一系列问题：已有成果的优化、单语种出版、网络化水平、类目体系问题等。除了依靠技术支撑之外,《中图法》本身还应积极采取对应措施：优化主类表结构、进一步修订复分表、规范类目描述、发行多种译本、继续推进《中图法》标准化等。第五部分,语义Web环境中KOS语义互操作与《中图法》SKOS描述转化。KOS要以数据关联形式在网络上发布和检索,实现语义互操作,必须采用标准的描述语言重新描述。SKOS语言是W3C设计的一种标准的编码语言,专门针对传统知识组织系统的计算机表示和机器可处理。KOS的现代化改革一直坚持以类表结构少改动的原则,这与SKOS的定位完全吻合,即只描述不改动。但在《中图法》等传统类表的SKOS语言全描述转化研究中,伴随概念框架完美设计的同时,类表的一些细节和一些特性却无法实现语义最低损耗的描述转换。既要保证SKOS的独立性,又要让类表的语义损耗最低、数据交互性最强大,办法之一是只转化类表前三级类目体系,以构建传统类表的顶层表示模型。之二,用SKOS全描述转化各类表,同时归纳各类表概念构建出顶层本体,在顶层本体数据与SKOS数据间建立联系。如此,在未来智能化的网络环境中提供真正的一站式网络检索服务的理想才能实现。更多还原

【Abstract】 For a long time, many experts and scholars of Library and Information Science are focused on the problem that the use of different kinds of retrieval languages in the same or different retrieval system leads to the diversification of retrieval strategy, and also has a bad effect on the sharing of information resources in the system. At the same time, these experts are also searching for the solutions. The standardization of retrieval languages is the best way, but it’s difficult to realize. The more realistic and feasible method is to establish the semantic mapping relation among the isomerism retrieval language vocabulary by manual or automatic operation, so that their compatibility and conversion can be realized, and finally that different kinds of retrieval languages can be integrated. In our country, Chinese Library Classification is the most effective general library classification with the largest user group. How to be compatible with other Chinese and foreign retrieval languages has received more attention. Only when it is integrated into this big system of retrieval languages can Chinese Library Classification keep its superiority status in knowledge organization systems. This paper, therefore, is divided into four parts to clarify the systemic research.The first part is about the brief introduction of related fundamental problems on the retrieval languages’compatibility. Firstly, discriminate the concept. On the foundation of tracing the concept of compatibility’ properties, process and transition, inter operability, mapping and converting etc., this part clarifies their relationship and formulates the concepts of retrieval language’s compatibility in both narrow sense and broad sense. Secondly, discuss the theoretical base of retrieval language’s compatibility from the aspects of consistency of basic principle and complementary of composition method about retrieval languages. Finally, analyze factors that influence the retrieval languages’compatibility:compilation principles of retrieval languages, class table vocabulary (category item concept) theme category, specificity, pre-coordinating, the Similarity on the class table structure etc.The second part is about the development of Chinese Library Classification’s compatibility. Specifically, it studies Chinese Library Classification absorbing and learning the ways of classification from home and abroad at the beginning of its establishment, Chinese Library Classification’s serialization table of classification and Classification-Subject Integration Thesaurus based on manual research, the research of Chinese Library Classification’s interoperability based on computer, the integration between Chinese Library Classification and other systems.The third part is about the compatible range and pattern of Chinese Library Classification. On the basis of comprehensively investigating and analyzing the achievements, research projects and plans of the compatibility between Chinese Library Classification and other retrieval language, this part points out that most of the Chinese Library Classification’s compatible achievements are the compatibility of its serialization table of classification and Chinese Thesaurus at present. Meanwhile, on the basis of Classified Chinese Thesaurus, the research of compatibility of Chinese Library Classification and natural language began. The compatibility of Chinese Library Classification and foreign classification language and subject language is largely limited to partial fields or subject and needs further study. Aimed at the mentioned compatible range, this part summarizes all kinds of features and technical properties of compatible formats which are used in the process of Chinese Library Classification’s realization of compatibility.The forth part is about the problems existing in the development of Chinese Library Classification’s compatibility in network environment, and also about the solutions. As the compatible range of some famous classified foreign language enlarges, the development of retrieval language’s compatibility becomes more and more important:organizing objects’expanding and further detailing, organization tool’s using pluralism, enlargement of concept system’s range, methods of strengthening and combining with traditional knowledge classification, various existence of knowledge organization tool, more and more obvious features of interdisciplinary development, etc. If the Chinese Library Classification wants to be a general retrieval language in Chinese circles and be compatible with other foreign retrieval languages, there is a series of problems:existing achievement’s optimization, publish of single language kind, networked level, problems of category system, etc. Besides depending on technical supports, Chinese Library Classification itself should take corresponding measures:optimize the structure of main schedule, further revise redistribution tables, standardize description of category, issue various translation versions, proceed promoting standardization of Chinese Library Classification, etc.The fifth part is about the semantic interoperation of KOS and description transformation research of Chinese Library Classification SKOS (Simple Knowledge Organization System) in semantic web environment. KOS have to use standard description language to describe renewably if it wants to publish and retrieve, and realize semantic interoperation on the internet as the form of Linked Data. SKOS language is a kind of standard coding language designed by W3C, aiming at computer representation and machine process of traditional knowledge organization system. Modernization Reform of Chinese Library Classification insist on the principle of few alterations of class table structure, coinciding to the orientation of SKOS, it is, description without alteration. But in the research of comprehensive description transformation of traditional class table, such as Chinese Library Classification, with the perfect design of concept frame,some details and characteristics of class table cannot realize the semantic description transformation with lowest loss. To guarantee the independency of SKOS and realize the lowest semantic loss and the most powerful data interaction of class table, one way is to transform the first three grade category system to construct traditional class table’s top-level representation model. The second way is to comprehensively describe and transform all kinds of class table by SKOS language, and induce their concepts to structure out top-level ontology, and establish contacts between top-level ontology data and SKOS data. So, the idea to provide real "one-stop" network retrieval services in the intellectualized network environment can come true in the future.更多还原