摘要
为提高编程题目的推荐效果与学习者学习效率,提出一种基于关联规则与知识追踪的编程题目推荐算法。结合学习者完成编程题目需经过多次尝试提交的特征与学习者遗忘特征,通过关联规则挖掘出学习者在同一知识点下同时做对或做错的题目关联特征,使用知识追踪模型与判别分析模型分析学习者对知识点的掌握状态,推荐巩固题、拔高题或下一知识点的题目。实验结果表明,该推荐算法比几种对比算法具有更高的准确率与覆盖率。
To improve the programming question recommendation effect and learners’learning efficiency,a programming question recommendation algorithm based on association rules and knowledge tracking was proposed.The characteristics that learners needed to submit after multiple attempts were considered to complete programming questions,and learner forgetting characteristics were combined.The relevant characteristics of questions that learners did right or wrong at the same time under the same knowledge point were mined by association rule.The knowledge tracking model and the discriminant analysis model were used to analyze learners’mastery of knowledge points and then to recommend consolidating questions,advanced questions or next knowledge points.Experimental results show that the proposed algorithm has higher accuracy and coverage than several compa-rison algorithms.
作者
向程冠
周东波
李雷
王英
XIANG Cheng-guan;ZHOU Dong-bo;LI Lei;WANG Ying(National Engineering Laboratory for Educational Big Data,Central China Normal University,Wuhan 430079,China;School of Mathematics and Big Data,Guizhou Education University,Guiyang 550018,China)
出处
《计算机工程与设计》
北大核心
2022年第11期3135-3142,共8页
Computer Engineering and Design
基金
科技创新2030新一代人工智能重大基金项目(2020AAA0108804)
国家自然科学基金项目(61977030)
国家自然科学基金项目(61807012)
教育部-中国移动科研基金项目(MCM20200406)
贵州省教育厅创新群体研究基金项目(黔教合KY字[2021]022)
贵州省科技厅国家自然科学基金奖励补助基金项目(黔科合平台人才[2018]5778-07)。
关键词
推荐算法
关联规则
知识追踪
判别分析
题目推荐
编程
学习效率
recommendation algorithm
association rules
knowledge tracing
discriminant analysis
question recommendation
programming
learning efficiency
