摘要
目前,自然语言处理系统由于缺乏语义信息及知识理解和推理能力,因此存在很多缺陷,在自动问答领域中很难分清用户表达的真正意图,并给出用户满意的答案.本文就此给出基于句法语义的问句分析方法,采用本体构建技术构建答案库,在此基础上进行领域知识语义推理,实现智能问答.本研究以黄山旅游为限定领域建立问答系统,试验结果表明多策略的方法是有效的.
Current natural language processing systems have some shortcomings because they lack knowledge understanding and reasoning abilities. These systems cannot catch the exact meaning of the users, thus they cannot meet their needs. Our research thinks out an approach based on syntactic and semantic analysis for questions analysis. An answering bank is established by means of ontology construction technique. On this basis, semantic reasoning and intellect answering can be achieved. Here, ‘Tour in Mount Huangshan＇ is chosen as the restricted domain, and a question-and-answering system is constructed. The experiment shows that the system is effective.
出处
《北方工业大学学报》
2010年第1期23-27,共5页
Journal of North China University of Technology
基金
北京市教委科研计划项目(KM201010009008)
北方工业大学校科研基金资助项目(英汉多词表达的抽取与语义解释研究)
北京市属高等学校人才强教计划资助项目(PHR201007121)
关键词
问答系统
本体构建
句法分析
question-and-answering system
ontology construction
syntactic analysis
