基于认知诊断的个性化试题推荐方法

发布时间：2018-03-02 11:37

本文选题：协同过滤　切入点：个性化试题推荐　出处：《计算机学报》2017年01期 　论文类型：期刊论文

【摘要】：面向学生的个性化试题推荐是智能教育领域重要的研究课题,现有的试题推荐工作大多采用协同过滤的方法或基于认知诊断的方法.然而,协同过滤的试题推荐方法往往忽略了学生的学习状态(知识点掌握情况);基于认知诊断的方法只能建模单个学生的学习状态,不能利用相似学生的共性特征.针对以上问题,文中提出一种基于学生知识点掌握程度的协同过滤试题推荐方法.该推荐方法分为3步:第1步结合认知诊断模型,根据学生已有的答题情况和试题知识点的关联对学生的试题掌握水平进行建模;第2步将学生的试题掌握水平用于概率矩阵分解预测学生的答题情况;第3步根据得分预测和试题难度向学生进行相应的试题推荐.该推荐方法同时考虑了被推荐学生学习的个性和群组学生学习的共性,在保证试题推荐解释性的同时提高了试题推荐的可靠性.最后,文中通过大量对比实验证明了该方法在进行学生试题推荐时能够保持精确性和可解释性.
[Abstract]:Individualized examination question recommendation for students is an important research topic in the field of intelligent education. Most of the existing work in this field is based on collaborative filtering or cognitive diagnostics. The collaborative filtering method often ignores the students' learning state (knowledge point mastery); the cognitive diagnosis based method can only model the learning state of a single student, and cannot utilize the common characteristics of similar students. In this paper, a collaborative filtering method for recommending test questions based on students' mastery of knowledge points is proposed. The method is divided into three steps: step 1, combined with cognitive diagnosis model, In the second step, the students' mastery level is used to predict the students' answer by probability matrix decomposition. In the third step, the students are recommended according to the score prediction and the difficulty of the test. The recommended method takes into account the personality of the recommended students and the commonness of the group students. At the same time, the reliability of the recommended questions is improved. Finally, through a large number of comparative experiments, it is proved that this method can keep the accuracy and interpretability of the students' questions recommendation.
【作者单位】：中国科学技术大学计算机科学与技术学院;合肥工业大学计算机与信息学院;安徽大学计算机科学与技术学院;科大讯飞股份有限公司;
【基金】：国家杰出青年科学基金(61325010) 国家自然科学基金(61403358) 中国科学院青年创新促进会会员专项基金(会员编号2014299) 安徽省自然科学基金(1408085QF110)资助~~
【分类号】：TP391.3

【相似文献】