Exploiting Term Importance Categories and Dependency Relations for Natural Language Search

摘要：

In this paper,we propose a method that clearly separates terms (words and de- pendency relations)in a natural language query into important and other terms,and differently handles the terms according to their importance.The proposed method uses three types of term importance:nec- essary,optional,and unnecessary.The importance are detected using linguistic clues.We evaluated the proposed method using a test collection for Japanese infor- mation retrieval.Performance was resul- tantly improved by differently handling terms according to their importance.

作者: Keiji Shinzato Sadao Kurohashi

作者单位: Graduate School of Informatics,Kyoto University Graduate School of Informatics, Kyoto University

会议类型: 国际会议

会议名称: The 23rd International Conference on Computational Linguistics(第23届国际计算语言学大会)

会议地点: 北京

会议语种:英文

页码: 2-11

在线出版日期: 2010-08-01（万方平台首次上网日期，不代表论文的发表时间）