基于知识图谱的智能答疑系统研究  

RESEARCH ON INTELLIGENT QUESTION ANSWERING SYSTEM BASED ON KNOWLEDGE GRAPH

作　　者：陈志云[1] 商月 钱冬明[1] 

机构地区：[1]华东师范大学,上海200062

出　　处：《计算机应用与软件》2018年第2期178-182,共5页Computer Applications and Software

基　　金：国家高技术研究发展计划项目(2013AA01A211)

摘　　要：在大型开放式网络课程(MOOC)系统中,答疑是一个重要环节,智能答疑系统的研发,有助于提高教学的效能。一般的答疑系统缺乏对课程知识的智能表示,对以往问题答案的自动归纳等。利用知识图谱技术将学生提问问题以知识点树的形式显示,同时结合朴素贝叶斯算法文本分类,将问题关键字按照知识点树中的知识点进行归类,并与教学案例课件进行分类匹配,从而初步改造为智能答疑系统。系统以华东师范大学公共计算机课为应用实例,初步应用测试问题分类成功率均在80%左右,证明智能答疑系统能更好地表示和分类问题,更好地适应学生的答疑应用。In the massive open online courses( MOOC) system,the question answering is an important link. The development of the intelligent question answering system is helpful to improve the teaching efficiency. The general question answering system is lack of the intelligent representation of the course knowledge,and the automatic induction of the previous questions,and so on. This paper used Knowledge Graph technology to represent questions in the form of knowledge point tree. At the same time,this paper classified problems according to knowledge points in the knowledge tree,and match problems with teaching case courseware by the Naive Bayes algorithm. The system,which initially transformed into intelligent answering system,took the public computer course of East China Normal University as an application example. After testing four classes of students studied Data Processing and Management,we found that classification success rate can be about 80%,which proved that intelligent answering system could represent and classified problems better,thus being suitable to Q & A for students.

关 键 词：中文知识图谱 智能答疑 朴素贝 叶斯问题分类 

分 类 号：TP311.53[自动化与计算机技术—计算机软件与理论]