基于贝叶斯网络的学习者学习行为评估算法研究  

Research on Evaluation Algorithm of Learner's Learning Behavior Based on Bayesian Network

作　　者：曾蔚[1,2,3] ZENG Wei(College of Mathematics and Computer Science,Quanzhou Normal University,Quanzhou 362000;Fujian Provincial Key Laboratory of Data Intensive Computing,Quanzhou 362000;Key Laboratory of Intelligent Computing and Information Processing,Fujian Province Univers让y,Quanzhou 362000,China)

机构地区：[1]泉州师范学院数学与计算机科学学院,福建泉州362000 [2]福建省大数据管理新技术与知识工程重点实验室,福建泉州362000 [3]智能计算与信息处理福建省高等学校重点实验室,福建泉州362000

出　　处：《太原师范学院学报（自然科学版）》2018年第4期29-34,共6页Journal of Taiyuan Normal University:Natural Science Edition

基　　金：2016年福建省中青年教师教育科研项目(JAT160419)

摘　　要：针对现有大部分教学平台提供给学习者相同的学习资源,无法对学习者的学习情况、知识点掌握情况进行实时评测和评估的情况,提出基于贝叶斯网络的学习者学习行为评估模型,主要针对学习者在教学系统中的解题过程行为进行评估及预测.文中首先提出的解决方法是对于每一类题目根据学习者的解题过程分别构建贝叶斯网络模型,但这种方法也存在着一定的不足之处.在此基础上,根据题目之间的关联性进而提出链式贝叶斯网络模型.最后采用KDD 2010比赛中提供的智能辅导教学系统学生日志数据分别验证这两种方法的可行性及有效性.Most of the existing teaching platforms which provide the same learning resources for the learners can't evaluate the learning situation and the mastery of knowledge of the learners in real time. According to this situation,a learner?s learning behavior evaluation model based on Bayesian network which mainly evaluates and predicts learnersy behavior in solving problems in Tutoring System was proposed. The first solution is to construct a Bayesian network model for each problem according to the learners? problem-solving process,but this solution has some shortcomings丒 On this basis,a chain Bayesian network model was proposed according to the relevance between problems丒 Finally, support for feasibility and validity of these two solutions was provided by applying them to logs of student interaction with Intelligent Tutoring System from the KDD Cup 2010.

关 键 词：贝叶斯网络 学习者学习行为评估 用户行为 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论] O235[自动化与计算机技术—计算机科学与技术]