国家语料库建设和汉语词表研究-计算机软件与理论专业论文.docxVIP

摘 摘 要 本文首先潮顾了语料库建设及语料瘁语言学的历史，介缁了国内外语料库建 设的现状，着黧介绍了国家语料库建设的基本情况，因为本文的工馋就是焱参与 潮家语料库建设课题的研究中完成的，这是本文开展研究的背景。 其次介绍了本人在站两年的主要研究工作——语料库的加工规范及词表的 鳐构纯夯辑工。邋两顼研究工{乍都是在联系簿筛麓汝占教授的汉语逻辫语义穰鍪理 论指导下进行的，与传统的汉语研究以及一般的中文信息处理研究都很不相同。 在滔表黪结掩纯掇工方瑟，挺密了缭稳纯谲表豹概念莠显开袋了工糕研究，对运 9万词的词表谶行了3次结构化加工，其结果已经运用在国家语料滕的加工实践 中，成效显著；在鸯曩王援范穷瑟，在结橡化调表熬基磷上提爨了结稳伲掭注鹁方 法，使得加工结果能够适应不同应用系统的需求。而且由于缩构化词表标注了许 多词法傣息，这藏使褥切词标注软件的设计修改与切词标注艘范之阙、_}霹袭与切 谰标注软件以及语料校对与最后输出结果之间都能保持相对的独立性，既减少了 人工校对的工作量，又保证了语料麾加工的一致性与正确率。 最后是未来工作的设想，提出了今瑶努力的其体方向。 关键词：潺辩瘁，蕊王麓范，结褥鬣丽表，良透滋辑语义模型 AbStraCt this Papelfirstly retrospects the history of corpus construction and COrl：}US linguistics．introduces the of COrpUS constructions home and abroad．The Paper focrises introducing the basic situation of National Corpus construction， since血e work this PaDer iS accomplished through the research of the task of National Corpus construction．This is the whole background this paper studies in。 Secondly,this Paper introduces the main postdoctora{research WOrk I have done during the two years：The Processing Criterion of Corpus and the Structural Processing of Word-stoc羹These two iterns was under the instruction of Prof．己0S Chinese Semantic Logic Model Thoery,SO it differs from the traditional Chinese study and general Chinese狐formafion Processing。In the aspect of the structural processing of Word-stock we brought forward the conception of structuralized 弭白rd．stock and started research from蜘e conception÷鹏have processed approximately 90000．word-Word-stock for three times using the structural way．And its result was used in the practice of恤e processing of national and got good effect．In the aspcct of Processing Criterion，we brouI曲t forward the method of stmctural tagging based the idea of stmcturalizcd Wbrd-stock,in order to make the processing result to fit the needs of different applications．Since we saved lot of loxical information in the structuralized Wo婚stock we make the criterion of tagging independent of the design and implementation of tagging software， wor