摘要： 句模匹配方法是基于字符表层的文本分析技术，只能处理各个关键词中有分隔词的问句，具有局限性。针对该问题，结合问答系统的应用背景，提出将句模匹配与关键词词库匹配相结合的方法，改进了传统的句模匹配问句理解方法。实际应用表明，该方法能扩大问句理解的范围，提高问句理解的准确性。

关键词: 问答系统, 问句理解, 自然语言处理, 句模匹配算法, 词库匹配算法

Abstract: Matching algorithm of sentence template can only be used in the sentences that have some divided words between the parameters, which is a technique for text analysis based on characters. So this method has its own limitations. In light of these problems and the background of question answering system, this paper presents a new method that combines with matching algorithm of lexicon, which improves the traditional method. In practice, result shows that the method can extend the range of question interpretation and enhance the accuracy of question interpretation.

Key words: question answering system, question interpretation, nature language processing, matching algorithm of sentence template, matching algorithm of lexicon

中图分类号: 

• TP311