基于贝叶斯网络的评估导学方法研究-计算机应用技术专业论文.docxVIP

基于贝叶斯网络的评估导学方法研究摘要 基于贝叶斯网络的评估导学方法研究 摘要 贝叶斯网络作为一种表达不确定性知识的有效工具，在许多领域 都得到了广泛的应用，本文将贝叶斯网络应用到智能评估导学系统 中，以实现对学生适应性评估和个性化导学。 本文首先对评估导学研究中学生模型的构建方法进行了比较分 析，在两种常见的贝叶斯网络学生模型基础上，提出了基于知识关系、 具有预测能力的覆盖型贝叶斯网络学生模型，通过量化知识项之间的 不确定相关性，使得该学生模型具有很强的预测能力，能够很好地反 映学生在特定领域内的知识结构。通过对学生模型特性的分析，本文 提出了基于自适应测试学生认知能力的评估方法，其中主要提出了自 适应选题算法，以确定适合学生认知水平的测试试题，并通过测试发 现学生的非真实能力，减少了不正常反应的干扰，提高了对学生认知 能力评估的准确度。进一步，本论文设计了相应的高层次教学策略和 低层次教学策略，以提供导学建议，实现对学生的个性化指导。最后， 本论文结合多Agent技术，在Jadex多Agent开发平台上开发了基于 贝叶斯网络的评估导学系统，并对在校学生进行了实际测试，测试结 果基本上符合学生的实际情况，而且给出了相应的指导意见。 关键词学生模型，贝叶斯网络，评估，导学，自适应测试 ， RESEARCH RESEARCH oN BAYESIAN NETWoRK BASED E、7ALUATING AND TUTORING APPROACH ABSTRACT Bayesian Network is a very effective tool to express uncertain knowledge and is used in many research fields widely．In this paper, Bayesian network is used in Intelligent Tutoring System in order to evaluate and instruct student individually Firstly,the student modeling approach is analyzed in this paper Based on two common Bayesian Network student models，an effective overlay student model with strong prediction ability based on domain knowledge relationship is proposed．This student model could reflect the domain specific knowledge structure and have strong prediction ability through quantiflying the uncertain relationship between knowledge items． Then，an adaptive evaluating method to test student’S cognitive ability is presented，in which the adaptive algorithm of selecting testing items is proposed to choosee the test questions fitting the student’S cognitive level． This evaluating method can decrease the disturbance of abnormal reaction and improve the exactness of evaluating student’s cognitive ability．Further more，a high—level teaching strategy and low—level teaching strategy has been provided in oreder to provide art individual instruction instruction to students．Finally,all evaluating and tutoring system based on Bayesian Network has been implemented under the multi-Agent platform Jadex．This system has been used to test some sophom