摘要智能授导系统ITS研究的重点在于增加远程网络教学系统的适
摘要
智能授导系统ITS研究的重点在于增加远程网络教学系统的适 应性和智能性。学生模型是ITS的基础和核心,是其他模块进行工 作的前提。论文针对目前国内ITS中学生模型构建偏于简单而缺乏 智能性的不足,结合人工智能领域的一些成熟技术,研究开发了一 个基于贝叶斯网络的智能授导系统AIBNS。
论文首先分析了基于前提关系和继承关系的两种建模方式的优 缺点,在划分知识项后,经过网络结构和参数学习,提出了一种构 建覆盖型贝叶斯网络学生模型的方法。通过对领域相关及无关信息 分别建模,构建出学生在特定领域内的知识结构,使模型具备较强 的预测能力。针对学生认知状态评估上存在的不确定性,以及测试 过程中不正常反应的干扰问题,论文提出了一种自适应选题算法, 以期用尽量少的试题完成对学生认知状态的精确评估,并减少评估
测试中的噪声。此外设计了一个基于模糊变换原理的学生认知状态 评估模型,不仅可以对相关知识项进行认知评估,而且可以进行多 级别的综合评价,提高了对学生认知状态评估的准确度。论文应用 贝叶斯网络面向对象的知识表达方式,实现了AIBNS原型系统,对 贝叶斯网络理论在ITS领域的实际应用提供了有益借鉴。
AIBNS系统区别于传统“以教为主”教学模式,实现了个性化 自主学习与交互协调学习相结合,不仅能够根据学生的不同个性自 适应地提供教学资源,而且可以针对学习目标提出建议,呈现给用 户学习知识的步骤序列。验证表明基于贝叶斯网络的智能授导系统 具有良好的智能性、交互性和适应性。
关键词智能授导系统,学生模型,贝叶斯网络,覆盖模型,模糊变
换
ABSTRACTThe
ABSTRACT
The focus of intelligent tutoring system(ITS)is tO increase the adaptiveness and intelligentness of the long—range network education system.The student model is the foundation and core of ITS,the prerequisite of performance of other models.The article demonstrates t11at,in view of present student modeling of ITS in China,which is to simple and lack of intelligentness.Combined with some mature
technology of artificial intelligence(AI),an ITS of址BNS based on
bayesian network(BN)has been developed.
This thesis first analyses the advantages and disadvantages of the
two kinds of modeling based on preparation relationship and inheritance relationship.After dividing knowledge item(I(I)and learning network
structure and parameter,the thesis puts forward a approach to form an
overlay student model based on BN,and after modeling on domain related knowledge and independent information.it shows the student’S knowledge structure in a specific field.which enables the model to have
more strong calculation ability.In view of the existing uncertainty of the evaluation of cognition of students and the disturbalice of abnormal reaction in calculation.it also gives out aIl adaptive item selection algorithm,which precisely evaluates cognition of students by using as few quest