MOOC在线学习行为的人类动力学分析

Human Dynamics Analysis on MOOC Online Learning Behaviors

近年来，随着互联网技术的深入推进，在线学习模式和MOOC 平台得到了日益普及和发展，与传 统课堂教学模式形成互补。同时，在线学习也得到了教育学界的普遍关注。为了深入了解用户在线学习的行为 模式和特点，本研究采用人类动力学研究方法，对国内MOOC 平台“学堂在线冶的用户学习抽样数据进行定量分 析发现，用户的在线学习行为具有明显的异质性(即非均匀性)，表现在:1) 用户的选课量和课程的选课人数有 很大差异;2)用户在线学习时间间隔分布和持续时间分布都表现为幂律分布，因而在线学习行为具有阵发和重 尾的特征;3)对用户学习行为的活跃性研究发现，用户学习的活跃天数大多较短，在线学习时间和次数也服从幂 律分布，且随着时间的推移，用户的平均学习时间会上升，在课程学习后期的讨论环节会投入更多时间和精力。

Online learning is such a developing field. utilizing learning platforms such as MOOCs， users can carry out learning at any time and any places. but due to the deficiency of supervision and motivation， lots of learners quit the learning process. Therefore， studies on behavior patterns of online learners can provide great help to further understand online learning behaviors， design learning platforms and make learning rules. the emergence of big data era provides us an opportunity to study online learning behaviors from a new perspective. Taking advantages of the complete and accurate dataset， we are able to explore deeper into learning mechanism and regularity. This paper conducted a quantitative analysis using the methodology of human dynamics， which is used to study the statistical property， scaling law and dynamical mechanism of human temporal-spatial behaviors. Via analyzing an open dataset from a Chinese MOOC platform， we found a heterogeneity or inhomogeneity in online learning behaviors， specifically from the following three perspectives:both the number of courses one user selects and the number of users who select one certain course show great discrepancy;the distributions of both the inter-event time and duration time of online learning follow power-law decay， indicating the burst and heavy tail property of online learning behaviors;the activeness of online learning behaviors shows a discrepancy between different users and different stages. Specifically， users spend different days and time on the learning process. Lots of users drop out from learning although the users who persist spend more time and put more effort on the courses The conclusions hopes can facilitate a better understanding of the statistical characteristics of online learning behaviors and provide explanations for people who drop out from learning.