基于正则等价的虚拟学习社区角色分类

Regular Equivalence-based Role Classification in Virtual Learning Community

研究者们常采用传统的机器学习方法在虚拟学习社区中提取以中心性或声望为主要标准的领袖节点。这些方法虽然简洁直观,但容易遮蔽虚拟学习社区的部分特点,忽视非领袖节点中也可能存在其他类型的“关键少数”。为了深入理解虚拟学习社区的人际关系网络,文章以某门慕课讨论区中26次讨论的86名学习者为例,采用基于正则等价的块模型方法,从中划分出4个不同角色,并解释了他们在知识构建过程中发挥的作用。结果发现:可根据知识传递的关系,将学习者的角色分为“导学者”“善学者”“熟练者”和“初学者”4类。这一结果不仅表明基于正则等价的虚拟学习社区学习者角色分类方法比传统机器学习分类方法(如K-means)划分出来的角色更细致,更能够发现直观视野之外的“关键少数”,还给虚拟社区的教学实践带来新的启发。它启发我们:如果能对不同的角色采用不同的教学策略,有可能进一步减轻社区助教的工作负荷,用更少的干预促进虚拟学习社区形成更浓厚的学习氛围。

Researchers often adopt traditional machine learning methods to identify leader nodes in a virtual leaning community based on centrality or prestige.These methods,while simple and intuitive,tend to obscure some characteristics of virtual learning communities and ignore the fact that there may be other types of"critical minorities"among the non-leader nodes.In order to gain a deeper understanding of the interpersonal network of virtual learning community,this paper takes 86 learners from 26 discussions in a MOOC discussion forum as an example.Using a block model approach based on regular equivalence,four different roles are identified and their roles in the process of knowledge construction are explained.The results show that the roles of learners could be classified into four categories,namely"guide learner","good learner","skilled learner"and"beginner",according to the relationship of knowledge transfer.This result not only shows that compared with traditional machine learning classification methods(e.g.K-means),the classification of learner roles in virtual learning community based on regular equivalence is more detailed,more able to find out the"critical few"beyond the intuitive view,but also sheds new light on the teaching practice in virtual community.It suggests that if different teaching strategies can be used for different roles,it may be possible to further reduce the workload of community teaching assistants and promote a stronger learning atmosphere in virtual learning communities with less intervention.