Abstract

Group formation is an important area of research given its application in online collaborative learning tools. Although several approaches have been proposed, they do not necessarily guarantee fixed-sized groups, a necessary condition to ensure that groups remain small. In this paper, a fixed-sized group formation method is proposed which utilises a modified k-means clustering algorithm in 2D feature space. Using an existing dataset of 34 learners, an experiment was conducted which clustered these learners into six groups, with a maximum of six learners in each group. These results were compared with the Random Group Formation and G2Group Approaches. The Fixed-Sized Group Formation out-performed Random Group Formation, but did not create as many compact clusters as G2Group. Importantly, the maximum size of each group was fixed and the generated groups were homogeneous. This method is important especially in applications where there is a need to create small collaborative groups, for example, in cross-classroom collaborative project-based learning.

Citation

Walcott, P.A. & Rolle-Greenidge, G. (2022). Fixed-Sized Group Formation Using a Modified K-Means Clustering Algorithm. In T. Bastiaens (Ed.), Proceedings of EdMedia + Innovate Learning (pp. 1-5). Online: Association for the Advancement of Computing in Education (AACE). Retrieved September 30, 2023 from https://www.learntechlib.org/primary/p/221261/.

© 2022 Association for the Advancement of Computing in Education (AACE)