摘要
以海量的问答记录为研究对象,采用问题语句的主题划分和类别划分相结合的分类方法对问题语句进行理解与组织,形成问答知识库的基本组织框架,基于此依据问答知识的使用频次将其划分为高频访问知识和非高频访问知识,构建一个基于问答知识库的自动问答系统。实验表明:问题语句的两步分类法能够促使问答知识的均衡分布,且问答知识的合理划分有助于改善问答系统。
Taking huge amounts of QA records as study objects, the paper proposes a question classification method combining theme and categorization to form a basic framework for QA knowledge base. In this framework, knowledge re- cords are divided into frequently used ones and non - frequently used ones, and an automatic question answering system is constructed. Experiments show that the new classification method can balance the distribution of QA knowledge, and help to improve the efficiency of the QA system.
出处
《现代图书情报技术》
CSSCI
北大核心
2012年第9期15-22,共8页
New Technology of Library and Information Service
基金
2011年度国家社会科学基金青年项目"联合虚拟参考咨询系统的知识库研究"(项目编号:11CTQ003)的研究成果之一
关键词
问答知识
知识库
问答系统
QA knowledge Knowledge base QA system
