基于依存搭配抽取技术的平面媒体语言监测研究  

Language Monitoring of Print Media Based on Dependent Collocation Extraction

作　　者：邵艳秋[1] 申资卓 刘世军 SHAO Yanqiu;SHEN Zizhuo;LIU Shijun(School of Information Science,Beijing Language and Culture University,Beijing 100083,China)

机构地区：[1]北京语言大学信息科学学院

出　　处：《山西大学学报（自然科学版）》2019年第3期526-533,共8页Journal of Shanxi University(Natural Science Edition)

基　　金：国家自然科学基金(61872402);教育部人文社会科学研究规划基金项目(17YJAZH068);中央高校基本科研业务费专项资金(18ZDJ03)

摘　　要：文章采用基于依存分析的搭配抽取方法,在对语料进行句法依存分析的基础上抽取搭配,用词语间存在依存关系来定义搭配中的“共现”,通过计算两个词的共现频率和互信息来抽取更加合理的搭配。将基于依存句法分析的搭配抽取技术应用于近70年的《人民日报》语言资源监测研究中,实验结果显示,相较于单独的词语监测方法,搭配具有丰富的语境信息,可以更好地反映语言的历时变化和社会生活的变迁。This paper adopted an effective method for collocation extraction,based on dependency parsing(DP),and by conducting syntactic dependency parsing on the corpus and combining with co-occurrence frequency and mutual information,we extract syntactic collocation reasonably and usefully.We apply the method of collocation extraction based on syntactic dependency parsing to language resources monitoring research,which involved the corpus of People’s Daily about 70 years.Compared with the monitoring method based on single word,collocation can provide more detailed context information and reflect the change of language and society more accurately.

关 键 词：搭配抽取 依存分析 语言监测 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]