【作者】 叶忠杰；

【导师】 陈庆章；

【作者基本信息】 浙江工业大学 ， 计算机应用技术， 2007， 硕士

【摘要】 答疑是教学过程不可或缺的一个环节,答疑效果如何很大程度上取决于该环节受益面和是否能满足学生个性化需求,为此很多教育工作者致力于借助网络环境实现智能答疑系统的研究。本文以课程知识本体为出发点,综合FAQ和文档库技术,构建和实现了一种释疑匹配率高,适应多种询问方式的网络化智能答疑系统。论文的主要工作包括:1.构建课程知识本体。按教学大纲的要求建立课程知识本体(以《局域网技术与应用》课程为例),以此为基础界定课程本体FAQ库的范围和内容,同时,也从课程知识本体中导出课程知识关键词库。2.建立问题理解数据库。为了正确理解用户的问题,系统建立了问题理解数据库,主要包括问句信号词、问题信号词、问句标准类型与结构等信息。在此基础上,对于FAQ库答疑采用知识关键词和问句类型匹配技术获取答案;对于FAQ库无法回答的问题,则转向课程文档库以文档相似度匹配方式查找可能的答案文档。3.构建了课程文档库。课程文档库是课程本体FAQ库的自然延伸,主要任务是叙述性问题的答疑,目的是提高系统的答疑能力。答案文档的搜索主要采用VSM文档相似度计算技术,这时将用户问题也作为文档处理。为了提高响应速度,系统对课程文档库进行了预处理,即预先提取问题信号词和按TF-IDF提取若干个课程知识关键词。4.实现了智能答疑系统。采用.Net技术实现了《局域网技术与应用》课程的智能答疑系统,系统包括自然语言提问、浏览和搜索等多种功能。试用结果表明,本文所用方法和技术是有效的,应用于课程答疑具有较好的效果。系统的主要特点有:1.结合基于本体FAQ库的准确、快速和文档库技术的全面性,使系统的答疑性能和效率达到较佳水平;2.以课程知识本体为出发点确定范围和知识关键词汇,抛开“大而不全”标准词典和分词技术,基本消除分词过程,杜绝分词歧义,提高答疑效率。更多还原

【Abstract】 Question Answering (QA) is an indispensable part of teaching process. The effect of QA is largely decided by its accessibility and how well it can satisfy students’ individual needs. Many scholars are committed themselves to the realization of network-based intelligent question answering system. Starting from course knowledge ontology, this paper, by using comprehensive FAQ an document database technology, aims to develop a network-based intelligent question answering system, which enjoys high match ratio and multiple enquiring methods.The contents of this paper mainly cover:1. A course knowledge ontology is established. The paper firstly establishes a course knowledge ontology in line with the course syllabus, Taking the course of LAN Technology and Application as example, based on which the range and content of course ontology FAQ database are decided, and course knowledge keywords database is obtained.2. A question understanding database is established. To assure correct understanding of users’ questions, a question comprehension database is established which is composed of question signal words, question leading words, and standard question types and constructions. Within the FAQ database, knowledge keywords and question type match technology is used to answer questions, while questions outside the FAQ database will be directed to the course database and possible answer document will be obtained by document similarity match.3. A course document database is established. A course document database, the natural extension of course ontology FAQ database, is designed to answer narrative questions and thus improves the completeness of system. The search of answer documents is realized through VSM document similarity computation technology and questions from users are processed as documents. In order to improve the search speed, the system pre-processes course documents, namely picking up question type leading words or picking up course knowledge keywords by pressing TF-IDF.4. A intelligent question answering system is realized. By using .Net technology, an intelligent question answering system for the course of LAN Technology and Application is constructed. The system is characteristics of a functional combination of natural language questioning, browsing and searching. Experiments show that the method and technology introduced in this paper are effective. Good results are achieved by using the system in question answering.The major advantages of the system include:1. The FAQ performance and efficiency are greatly improved through course ontology FAQ database’s accuracy, quickness and the completeness of document database technology.2. Search range and knowledge keywords are determined from course knowledge ontology, discarded the bulky but incomplete standard lexicon and words segmentation technology, basically eliminated the words segmentation process and avoided words ambiguity. The overall question answering efficiency is improved.更多还原