Abstract:

The problem of learning loneliness is one of the reasons for the low completion rate of MOOC. Constructing a collaborative learning group that adapts to the learning characteristics of learners can solve this problem. This method uses auto-encoder to extract the key features of learners, and then uses the fuzzy C-means algorithm to iteratively group learners based on the principle of homogeneous grouping, so that online learners change from learning alone to collaborative learning. Thereby it improves online learning experience and reduces learning loneliness. Grouped online collaborative learning with 19,846 online learners who chose the introductory courses in computer science and programming on edX as experimental subjects. The results show that based on this method, learners in each group have high homogeneity, which can solve the problem of learning loneliness well.

Key words: learning loneliness, online collaborative learning group, auto-encoder, fuzzy C-means, MOOC

摘要：

学习孤独感问题是造成MOOC课程学习完成率低的原因之一，构建与学习者学习特征相适应的协作学习小组，可以有效解决学习孤独感问题。利用自编码神经网络提取在线学习者的关键特征，根据同质分组原则，利用模糊C均值算法对在线学习者进行迭代分组，使在线学习者从独自学习转变为以团队的形式进行协作学习，从而改善在线学习者的学习体验，降低学习孤独感。以edX平台上选择计算机科学与编程入门课程的19?846名在线学习者为实验对象，进行在线协作学习分组。实验结果表明，基于该分组方法，每个小组内学习者都有较高的同质性，可以很好地解决学习孤独感问题。

关键词: 学习孤独感, 在线协作学习分组, 自编码神经网络, 模糊C均值, MOOC课程