A Study of Text Topic Modelling Based on Role Dynamic Transformation in Online Collaborative Learning Platforms

With the rapid development of information technology, online collaborative learning has gradually become an important part of the educational field. This study focuses on the relationship between the dynamic transformation of learners' roles and text topics in online collaborative learning platforms. Taking the course of Instructional System Design in the College of Education of Zhejiang University of Technology as the research object, we collected discussion data from 120 students in synchronous (Wechat) and asynchronous (Learning Tong Forum) chatting environments, and analysed them by using BTM (Biterm Topic Model) and BERTopic Topic Modelling techniques for analysis. The results of the study show that there are significant differences in the distribution of text topics between synchronous and asynchronous environments, with the discussion topics in synchronous environments being more extensive but less in-depth, and the discussion topics in asynchronous environments being more focused and in-depth; the learner roles in both environments show a dynamic transformation, and the grouping mode with complementary personalities can help to enhance the efficiency of collaboration. Based on this, this study proposes practical suggestions for optimising online collaborative learning environments, including strengthening discussion depth guidance, optimising platform functions, and considering personality differences in grouping strategies, etc., with the aim of providing theoretical support and practical guidance for online collaborative learning instructional design and management.