Leveraging the Louvain algorithm for enhanced group formation and collaboration in online learning environments

Abstract

Abstract This study explores the dynamics of student interaction networks within an online asynchronous discussion forum, focusing on both whole group discussions and subgroup interactions distinguished by the Louvain algorithm, a renowned community detection method. Analyzing 2481 posts from 154 undergraduate students across three sections of a communications course centered on discussions about movie clips or social phenomena to enhance media literacy, this research aims to interpret the interaction patterns in these virtual spaces. Traditional methods of group formation, such as teacher intervention and self-selection, often fail to create balanced and effective groups, especially in large online courses. The Louvain algorithm, known for its efficiency in modularity optimization, identifies clusters based on actual student interaction patterns. By leveraging both global and local network analyses, this study provides a comprehensive understanding of interaction structures. The global network analysis offers a macro view of overall interaction structures, revealing diverse patterns despite identical course designs, suggesting the influence of unique group dynamics. The local analysis, focusing on the intricacies of node and edge connections, underscores that the Louvain algorithm's classifications correlate with heightened cohesiveness and collaborative potential. The results indicate that algorithmically detected groups exhibit strong internal communication and cohesiveness, as evidenced by high clustering coefficients, density values, and weighted degrees. These findings underscore the potential of network analysis to optimize online student interactions, providing valuable insights for refining educational design to promote student engagement and collaborative problem-solving. This research highlights the transformative potential of integrating advanced data-driven techniques in educational technology to improve group formation and collaborative learning outcomes, offering empirical insights for educators to enhance online interactions and expand pedagogical understanding.

Keywords

• Louvain algorithm
• Group formation
• Asynchronous discussion
• Cohesiveness
• Network analysis
• Community detection