基于Q/A与文档库的智能课程答疑系统  

Answering system of intellectualized curriculum based on Q/A and document

作　　者：叶忠杰[1] 陈庆章[2] 

机构地区：[1]浙江工业大学职教学院 [2]浙江工业大学信息工程学院,浙江杭州310032

出　　处：《计算机工程与设计》2008年第10期2692-2694,共3页Computer Engineering and Design

基　　金：浙江省新世纪教改基金项目(yb05103)

摘　　要：评价智能答疑系统优劣的重要指标是准确率和召回率。系统结合Q/A库和文档库搜索技术的优势实现,利用成熟的Q/A技术回答常见问题,保证了系统的准确率和高效率。利用智能文档搜索技术解答非常见问题,提高了系统的召回率,又因为事先对文档作了预处理,使搜索效率明显提高。同时系统基于课程开发,关键词的词汇量少而精确,使得语义理解的处理得以简化。The key criterion to evaluate the advantage and disadvantage of intellectualized answering system is its accuracy and recalling. If systemically combining the searching techniques between Q/A storehouse and document storehouse and put them into effect, also making use of adept Q/A technique to answer common questions, then it can ensure the accuracy and high efficiency of the system. The intellectualized document searching techniques is used to solve unusual problems, therefore the recalling of the system will be increased. Since pretreatment is taken for documents in advance, the searching efficiency obviously raised. Meanwhile, on the basis of developing curriculum, less and accurate key vocabulary, the dealing with understanding semantics becomes simplified.

关 键 词：智能答疑 召回率 向量空间模型 文档预处理 文档相似度 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]