汉语框架语义网中词元语义搭配模式确定规则研究  

Research on determining rule of lexical semantic collocation patterns in Chinese FrameNet

作　　者：余小鹏[1] 徐健儿 王振佩 姚小桐 YU Xiaopeng;XU Jianer;WANG Zhenpei;YAO Xiaotong(School of Management,Wuhan Institute of Technology,Wuhan 430205,China;School of Computer Science and Engineering,Wuhan Institute of Technology,Wuhan 430205,China)

机构地区：[1]武汉工程大学管理学院,武汉430205 [2]武汉工程大学计算机科学与工程学院,武汉430205

出　　处：《智能计算机与应用》2024年第8期29-31,共3页Intelligent Computer and Applications

基　　金：教育部人文社会科学研究规划基金项目(19YJA880077)。

摘　　要：机器理解自然语言和抽取信息涉及对词汇语义信息和句法信息两项重要内容的理解。汉语框架语义知识库(CFN)的词元语义搭配模式记录了词元与框架元素、词元与句法的组合方式,为机器自动进行语义分析和信息抽取提供了资源。而目前的词元语义搭配模式的数量多,机器确定并应用词元语义搭配模式存在一定困难,且框架元素间的关系比较模糊,词元配价成分的语义信息存在歧义,使机器不能准确理解和抽取文本信息。采用基于规则的方法,根据词元配价成分的语义信息和语义特征、词元外部的语境和框架元素提炼规则,可以使机器能够更准确地理解与抽取文本的语义信息。Machine understanding of natural language and information extraction involves the understanding of lexical semantic information and syntactic information.The lexical semantic collocation patterns of Chinese FrameNet record the combination of lexical units and frame elements,and lexical units and syntax,which provide resources for automatic semantic analysis and information extraction.There are a large number of lexical semantic collocation patterns now,which is difficult for the machine to determine and apply,and the relationship between frame elements is vague,and the semantic information of valency components of lemma is ambiguous,which makes the machine cannot accurately understand and extract text information.According to the semantic information and semantic features of the valence components of the lexical units and the external context of the lexical units and frame elements,the rules are extracted to make the machine understand and extract the semantic information of the text more accurately.

关 键 词：语义知识库 框架语义 规则 词元语义搭配 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]