个人信息Personal Information
副教授
博士生导师
硕士生导师
性别:女
毕业院校:大连理工大学
学位:博士
所在单位:计算机科学与技术学院
学科:人工智能
办公地点:大连理工大学创新园大厦B911
电子邮箱:zhouhuiwei@dlut.edu.cn
A Dependency Analysis Method for Japanese Sentences with Complex Conjunctive Relations
点击次数:
论文类型:会议论文
发表时间:2008-01-01
收录刊物:CPCI-S、CPCI-SSH
页面范围:184-187
关键字:dependency analysis; conjunctive relation; conjunctive tree
摘要:In Japanese, it is difficult to analyze a long sentence since it often has conjunctive relations. Japanese conjunctive relations are very complex. The structural ambiguity is common for Japanese sentences with complex conjunctive relations. This paper presents a Japanese dependency analysis method that first detects complex conjunctive relations and then analyzes the dependency relations by using the information about the conjunctive relations. The conjunctive relations can divide a long sentence into several shorter forms to reduce the analysis complexity. Experiments using the Kyoto University Corpus show that the proposed method improves the conjunctive relation accuracy and the dependency accuracy effectively.