【作者】 张小勤；

【导师】 滕至阳；

【作者基本信息】 东南大学 ， 计算机软件与理论， 2004， 硕士

【摘要】 答疑解惑是教学过程中不可缺少的环节,如何在Internet环境下设计具有自动答疑功能的网络自动答疑系统,成为当前智能教学研究的热点。目前国内外在这方面做的工作主要集中在如何提高系统的智能方面,代表性的算法有:关键词匹配算法、加权关键词算法等。这些算法具有一定的智能,但就返回答案的准确率以及系统的个性化服务等方面还存在很多问题。作者针对这些情况,做了以下两方面的工作:1应用模糊数学于网络自动答疑系统中,在分析关键词对于句子的隶属度、用户问题与标准问题的贴近度基础上,提出了基于阈值的择近原则算法。该算法克服了目前国内外几个代表性算法的不足,提高了返回答案的准确率,并且为用户提供了个性化的服务。2采用三层体系结构架构系统,设计并完成了基于J2EE体系结构的网络自动答疑系统NAAQS。通过对系统运行情况的分析,证实了基于阈值的择近原则算法具有高智能性。更多还原

【Abstract】 Question Answering is an important educational activity. How to develop an automatic answer question system running on the Internet, become a hotspot in the intelligent teaching area. Current works are focusing on how to improve the intelligence of the system, and there are some representative algorithms about it, such as: Based on the keywords matching algorithm and Based on the values keywords algorithm, etc. Those algorithms have a certain intelligence, but they also have some flaws on the correctness of answer and poor quality of service to the user. Toward this situation, two main works of this paper are as the following:1 applying the Fuzzy Mathematics in the system, this paper,based on the grade of membership between keyword and sentence, the proximity between user question and standard question, provides a new algorithm of the Full Close Principle with threshold, and it improves the intelligence of system.2 develop a Network Automatic Answer Question System(NAAQS) running on the Internet.,更多还原