基于问句语义表征的中文问句相似度计算方法  

Calculation Method of Chinese Question Semantic Similarity Based on Question Semantic Representation

作　　者：陈康[1] 樊孝忠[1] 刘杰[1] 贾可亮[1] 

机构地区：[1]北京理工大学计算机科学技术学院,北京100081

出　　处：《北京理工大学学报》2007年第12期1073-1076,共4页Transactions of Beijing Institute of Technology

基　　金：国家教育部高等学校博士学科点专项科研基金资助课题(20050007023)

摘　　要：提出了一种中文问句语义相似度计算的新方法.该方法分为两步:第一步采用基于问句句型模板规则匹配的方法提取问句语义表征;第二步根据问句语义表征计算问句语义相似度.采用该方法开发了一个面向常问问题集(FAQ)的问答系统.实验结果表明,采用该方法获得的相似度计算的准确率约为85%.A new approach to calculate the Chinese question semantic similarity is presented, that is divided into two steps, first to extract the semantic representation from the question, and secondly to compute the semantic similarity based on the question semantic representation. The method of extracting the question semantic representation from the question based on question semantic model matching is used. A FAQ oriented question answering system was developed based on the question semantic similarity calculation approach. Experimental results showed that the proposed algorithm has a higher precision, approximately 85 %.

关 键 词：问句语义表征 语义相似度 中文问句 语义块 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]