ISSN 1000-1239 CN 11-1777/TP

计算机研究与发展 ›› 2015, Vol. 52 ›› Issue (3): 614-628.doi: 10.7544/issn1000-1239.2015.20140491

• 信息处理 • 上一篇    下一篇

基于MOOC数据的学习行为分析与预测

蒋卓轩,张岩,李晓明   

  1. (北京大学网络与信息系统研究所 北京 100871) (jzhx@pku.edu.cn)
  • 出版日期: 2015-03-01
  • 基金资助: 
    基金项目:国家“九七三”重点基础研究发展计划基金项目(2014CB340405);国家自然科学基金项目(61272340,61472013)

Learning Behavior Analysis and Prediction Based on MOOC Data

Jiang Zhuoxuan, Zhang Yan, Li Xiaoming   

  1. (Institute of Network Computing and Information System, Peking University, Beijing 100871)
  • Online: 2015-03-01

摘要: 随着近2年慕课(massive open online course, MOOC)的兴起,教育大数据分析正成为一个新兴的研究方向.2013年秋,北京大学在Coursera上开设了6门慕课.通过分析挖掘约8万多人次参与这6门课的海量学习行为数据,力图展现慕课学习活动多个侧面的风貌.同时,首次针对中文慕课中学习行为的特点,将学习者分类,以更加深入地考察学习行为与学习效果之间的关系.在此基础上,通过选择学习者的若干典型行为特征,对他们最后的学习成果进行预测的工作也尚属首次.数据表明:基于学习行为的特征分析能有效地判别一个学习者能否成功完成学习任务获得通过证书,并能找出潜在的认真学习者,这为今后更加精准的慕课教学测评提供了一种依据.

关键词: 慕课, 学习者类型, 学习行为, 数据分析, 成绩预测

Abstract: With the booming of MOOC (massive open online course) in the past two years, educational data analysis has become a promising research field where the quality of teaching and learning can be and is being quantified to improve the educational effectiveness and even to promote the modern higher education. In the autumn of 2013, Peking University released its first six courses on the Coursera platform. Through mining and analyzing the massive data of learning behavior of over 80000 participants from the courses, this paper endeavors to manifest more than one side of learning activity in MOOC. Meanwhile, according to the characteristic of learning behavior in Chinese MOOC, learners are classified into several groups and then the relationship between their learning behavior and performance is thoroughly studied. Based on the above work, we find out that learners performance, regarding whether heshe could get certificated eventually, can be predicted by looking into several features of their learning behavior. Experiment results indicate that these features can be trained to effectively estimate whether a learner is probably to complete the course successfully. Besides, this method has the potential to partially evaluate the quality of both teaching and learning in practice.

Key words: massive open online course (MOOC), engagement style, learning behavior, data analysis, performance prediction

中图分类号: