季峰






摘 要: 为了提高试题个性化推荐系统对语法试题推荐的有效性,定义了韩语语法点矩阵、得分模型、潜在学习特征向量等,建立了学习态势算法,设计了试题智能推荐的整体框架。该系统以学习者的态势计算为基础,对学习者语法学习能力估算,再进行试题推荐。通过两组学习者对比,根据系统的平均使用时间、成绩提高率、答卷正确率和试题推荐精准率等指标进行评估,结果显示,该推荐系统对学习者有显著的助学效果。
关键词: 试题; 推荐系统; 语法; 学习态势; 框架
中图分类号:TP399 文献标识码:A 文章编号:1006-8228(2021)12-10-04
Abstract: In order to improve the effectiveness of the personalized test question recommendation system for grammar test question recommendation, Korean grammar point matrix, score model and potential learning feature vector are defined, the learning situation algorithm is established, and the overall framework of test question intelligent recommendation is designed. On the basis of learners' situation calculation, the system estimates learners' grammar learning ability, and then recommends test questions. Through the comparison between the two groups of learners, the evaluation is carried out according to the indicators such as the average use time of the system, the improvement rate of scores, the correct rate of answers and the accuracy rate of test question recommendation. The results show that the recommendation system has a significant help learning effect on learners.
Key words: test question; recommendation system; grammar; learning situation; framework
0 引言
全球知識一体化的教育框架下出现了很多语言类教育的创新,努力构建高沉浸感的语境,在语言类学习系统中,面对等级考试中的大量语法测试题,如何快速的提高语言水平,精准练习和消除冗余训练成为焦点,如何打破众多语言教育专家认为语法学习中需要要反复练习的经验思维,成为重要的课题,尽量减少练习周期掌握语法点显得尤为重要。
1 相关工作
国内外有不少团队致力于试题推荐领域的研究,王超等基于DINA模型提取个性化学习能力向量[1],建立多维离散认知模型,将试题推荐应用在几何学习中。赵泉通过计算知识点权重[2]结合课程需求建立推荐算法。朱天宇等基于协同过滤和认知诊断的推荐方法[3],引入个性化因子提高推荐精度。陈达基于内容相似度的最近邻方法[4],分别建立了基于内容和协同过滤的推荐系统。……