基于汉英词性组合的短语级平行语料类别知识挖掘研究  

Research of Mining the Category Knowledge Based on Chinese-English Part of Speech Sequence Parallel Corpus in Phrase Level

作　　者：王东波[1] 韩普[2] 沈耕宇[1] 沈思[3] 

机构地区：[1]南京农业大学信息科学技术学院 [2]南京邮电大学经济与管理学院 [3]南京大学信息管理学院

出　　处：《图书情报工作》2013年第11期106-111,145,共7页Library and Information Service

基　　金：国家自然科学基金面上项目"面向知识服务的知识组织模式与应用研究"(项目编号:71273126);国家社会科学基金重点项目"人文社会科学汉英动态术语数据库的构建研究"(项目编号:11AYY002)研究成果之一

摘　　要：基于通过具体实验确定的Bisecting K-means聚类和Lemmatization形态变换算法,在汉英短语级人文社会科学平行语料基础上,尝试进行类别知识挖掘的实验。在中文社会科学引文索引(CSSCI)的类别和标题知识基础上,完成对汉英语料的预处理,并分析名词、动词和形容词的分布状况。在名词、动词和形容词等词性的组合基础上,对比不同词性组合的效果并确定最优的词性组合类别知识挖掘模型。I The paper attempts an experiment of mining the category knowledge from Chinese - English humanities and social sciences parallel corpus in phrase level based on clustering and morphological conversion algorithms of Bisecting K - means and Lemroatization, which are determined by the experiments. The Chinese and English corpus preprocessing is completed and the distribution of noun, verb and adjective is analyzed based on Chinese Social Sciences Citation Index （CSSCI）. The results of different part of speech sequences are compared and the most optimal part of speech sequence applied to mining the category knowledge is determined on the basis of part of speech sequences of noun, verb and adjective.

关 键 词：词性组合Bisecting K-means汉英平行语料库 类别知识 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]