在线学习行为与学习效果——基于学习分析的实证研究
Online Learning Behavior and Learning Outcomes

网络教学平台凭借独特的时空便利性受到越来越多人的关注，通过挖掘学习者在技术平台上留下的海量行为数据，可以掌握学习者的学习状况，从而对学习过程起到干预和指导作用。鉴于此，文中采用了意大利热那亚大学计算机工程专业大一学生参加“数字电路”实验课程的行为数据集，通过对学生进行成绩预警分析，探索了学习行为与学习效果之间的关系。结论表明：在对学生能否通过最终测验进行不及格预警分析时，用分类树模型于课程学习的中后期阶段进行预测效果较好；在对学生的最终测验成绩进行分数预警分析时，用回归树模型于课程学习的前中期阶段进行预测效果较好。
Network teaching platform has attracted more and more attention due to its unique spatial and temporal convenience. By mining massive behavioral data left by learners on the technology plat-form, learners’ learning status can be mastered, thus playing an intervention and guidance role in the learning process. In view of this, this paper adopted the behavior data set of the freshmen ma-joring in computer engineering from the University of Genoa, Italy, who took part in the experiment course of “digital circuit”, and explored the relationship between learning behavior and learning effect by analyzing the students’ performance warning. The conclusion shows that in the early warning analysis of whether the students can pass the final test, the effect is better when the classification tree model is used to predict the grades in middle and late stage of the course learning. In the warning analysis of how many points can students get, the effect is better when the re-gression tree model is used to predict the grades in preintermediate stage of the course learning.