学习者知识模型的在线学习算法

Online learning algorithm for learner knowledge model

学习者知识模型是智能授导系统(ITS)中教学过程实现和策略实施的基础,然而由于判别学习者知识掌握程度的不确定性和学习者知识掌握水平的实时变化,构建能正确反映学习者知识掌握程度及其变化的知识模型十分困难。基于贝叶斯网络,以知识项为基本节点构建学习者知识模型的结构;引入问题节点,根据学习者的学习测试结果,采用Voting EM算法来对知识模型的参数进行在线学习和更新;同时,通过设置置信因子和更新时间标记来改进在线学习的效果。实验表明,模型能够较好地反映学习者知识掌握状况和快速适应学习者知识掌握水平的变化,有助于ITS更好地评价学习者学习效果。

Learner knowledge model is the foundation of the teaching process and strategy on Intelligent Tutoring Systems （ITS）. Because of the uncertainty of recognizing knowledge level of learner and the real-time changes of knowledge level, it is very difficult to construct a model to reflect learner＇s knowledge level and its change correctly. The paper used Bayesian network for learner knowledge modeling. According to knowledge level＇s change during learners＇ learning process, problem knot was introduced into knowledge model, and Voting EM algorithm was used for online learning and updating of knowledge model＇s parameters. Finally, the paper introduced confidence factor and time updating mark to improve the efficiency of online parameters learning and revise the result. The experimental results indicate that the model can reflect learner＇s knowledge status better, and can quickly keep up with the change of knowledge level. It can help ITS to evaluate the learning effects better.