问句相似度计算是基于常问问题库的问答系统的重点.现在的问句相似度计算准确率较低,为此,提出了一种基于主题和焦点的中文问句相似度计算方法.主题和焦点能够反映问句的主旨,识别出问句的主题能够更好地理解问句.其中抽取问句主题和焦点的方法能获取部分语义信息,而且比传统的根据疑问词进行语义分析的方法适用类型更广,同时在计算问句相似度时考虑了主题和焦点的影响.最后通过设计实验与其他方法进行比较,实验表明,该方法提高了准确率.
Sentence similarity computing is an important part in question answering system.The accuracy of the existing sentence similarity algorithm needs to be improved.An new method based on theme and focus of an question was presented.Theme and focus can reflect the purport of a question,and identify that it can better understand the question.The method extracting theme and focus can obtain some semantic information,it can be suitable to more question type than the traditional methods based on interrogative.It considers the impact of the theme and focus in questions similarity computing.At last,by designing experiment to compare with other methods,the experiment shows this method improves the accuracy.