【作者】 张虹；

【导师】 刘磊；

【作者基本信息】 南京农业大学 ， 图书馆学， 2008， 硕士

【摘要】 文献分类法的兼容互换已成为情报检索语言研究中亟待解决的问题之一。实现国内不同领域应用范围较为广泛的几种文献资料分类法的兼容互换,能够消除浏览、标引、检索中的障碍。实现文献分类法兼容互换能够满足信息检索和信息组织的迫切需要,更大程度实现文献信息资源的共享。文中以《中国图书馆分类法》(第四版)、《中文新闻信息分类与代码》和《社会科学检索词表》教育类为例分析了文献分类法兼容互换的可行性、实现原理,通过分析三部分类法类目之间的差异,认为类目之间应该存在多种转换方式。总结前文所述的国内外研究方法和经验的基础上,采用基于集成词表模式和对照索引模式的方法实现文献分类法兼容互换。文中把集成词表也看成广义上的对照索引,其中集成词表是文献分类法分类表结构和类目体系之间互相对照;其二是两种分类法的索引之间的对照,通过对照实现语义及语词字面的兼容互换。本文根据实际需要采用人工和机器结合的方法,在运用中对两种方法取长补短,主要运用词汇相似度计算的方法进行类目结构转换和字面转换,得到不同文献分类法的类目匹配对应结果,即是通过计算类名词、注释词、上位类类名词、下位类类名词以及类目对应的主题词等词汇的语义相似度,得到类目的相似度,进而确定类目之间的对照关系。在借鉴国内外分类语言互操作的经验和方法的基础上,设计和构建了一个可以方便修改和扩充的分类集成词对照表。本文中运用了Access2003、Visual Basic语言、Deamweaver等工具,采用计算机机械运算结果和人工相结合的方式构建一个具有多效能的实验系统,该系统能选择原有分类法单独查找,也能选择其中的两部分类法进行语词相似度匹配。实现了新闻领域与普通领域的不同分类法中具有相同语义的类目之间相互调用。文中参照《中图法》(第四版)编制了分类对照索引,即将不同词表的索引数据合并,利用计算机辅助编制轮排索引,此索引丰富了标引人员和用户查找、利用分类表和分类目录的方式,方便了分类法的使用。最后提出了有关分类法兼容互换的建议。更多还原

【Abstract】 With the development of Internet, especially library classification, the interoperability becomes an evergrowing demand in Internet searching. Users urge a system to unify various all kinds of Literature information, which may eliminate browsing, indexing and retrieval of obstacles.Compatibility and conversion of Literature Classification could meet the urgent needs of information retrieval and information organizations, and achieve a greater degree of information resources sharing information. This paper analyses the main methods of the cross-browsing and cross-searching based on the Multiple thesaurus between Chinese Library Classification and other classification schemes, including the choice of general classification scheme, the process of making Multiple thesaurus etc, also introduces the main function of Multiple thesaurus and library Literature of this project. On basis of concluding described the research methods and experiences at home and abroad. In this paper, the term sheet to achieve compatibility exchange of literature based on integrated thesaurus model and index model method. This paper in the form of the integrated thesaurus is also seen as the broadest sense of control index, which is matching each other between literature classification categories and classification structure, and the second is the two index of the classification, through mapping to achieve the compatibility and the conversion of Semantic and words literally through mapping. Based on the actual need for a combination of machines and artificial methods, the use of the complementarities of the two methods, mainly the use of vocabulary similarity calculation method archive structural transformation and literal conversion, corresponding match results different the literature classification categories, That is, by calculating semantic similarity of Term, including class terms, the word Notes, the upper class category term, the term under-class category and the corresponding categories such as MeSH terms, by similarity of categories, thereby determining the compatibility and conversion of the relationship between categories. This paper Construction and design on integration of the classifications can easily modify and expansion of the word tables. This paper in the use of the Access2003, Visual Basic language, and other tools, using computer and mechanical build a multi-effectiveness of the experimental system, the system can choose a separate classification of the original find, can also select the two of the Classifications to match words similarity, To achieve the call the different areas in the same categories of semantics between the news field information and the general classificationIn this paper, the light of "Chinese Library Classification" (fourth edition) prepared a classification mapping index, that is consolidation different thesaurus index data, the use of computer-aided round Index, the index has enriched the indexing and users find, use tables Categories and the way to facilitate the use of the classification. At last, it proposes some suggestions about the application of compatibility and conversion of Literature Classification.更多还原