中文问答系统中问题理解的研究与实现

Research and Implementation of Question Comprehension in Chinese QAS

围绕中文问答系统中的问题理解技术,本文研究了如何进行词法分析、问题分类、关键词提取及扩展、句型分析和浅层语义分析,主要提出了基于启发的疑问词和疑问焦点相结合的问题分类方法、问句统一型的句型分析方法和基于语义角色标注的语义分析方法。在此基础上,在Visual C++ 6.0环境下开发了一个基于简单的事实类问题的中文问答系统原型。实验结果表明:本文提出的方法对改善中文问答系统的性能是有效的。

The issue of question comprehension in Chinese question answering system (QAS) is discussed in this paper. The authors study word segmentation,question categorization,Keywords extraction and extension,sentence pattern analysis and shallow semantic parsing. The rules combined the question words with question focus based on heuristic to classify questions are introduced. As a result,a question pattern analysis method named question sentence standard formats and a semantic analysis method based on semantic role labeling are proposed. On the basis of the theory study above,a Chinese QAS prototype based on simple facts was developed using language Visual C++ 6.0. The experiment results show that the methods proposed in this paper improve the performance of Chinese QAS effectively.