摘要
FAQ(Frequently Asked Question)在互联网站上广泛使用,但绝大多数FAQ的检索与回复都是手工进行.本文介绍了一个较为完整的基于语句相似度计算的FAQ自动回复系统,包括基本计算模型的选取、FAQ特性的分析、FAQ数据形式化表示及特征向量索引、权重计算等,并给出详细的语句相似度计算算法,实验结果表明,对于频率高、共性大的问题,系统有很高的准确率.
FAQ（Frequently Asked Question） is widely used on the Internet, but most FAQ＇s asking and answering were not automatic. This paper introduces the design and implementation of a FAQ automatic answering system, including computation modal choosing, FAQ characters analyzing, FAQ data formal expressing, feature vector indexing, weight computing etc. , and gives a detail algorithm of similarity computing for question sentences. Experiment shows that the system has a good performance for high frequent and common questions.
出处
《小型微型计算机系统》
CSCD
北大核心
2006年第4期720-723,共4页
Journal of Chinese Computer Systems
基金
国家自然科学基金项目(60272088)资助
