An Approach to Group Formation in Collaborative Learning Using Learning Paths in Learning Management Systems

Forming groups in distance education is challenging for teachers because, with this modality, only 20% of the classes are held in person with the students. Thus, it is essential to achieve satisfactory results with automated approaches that can help teachers. In this article, an automated approach is proposed to assist teachers in recommending groups of students to learning management systems. We developed and validated a conceptual framework for group recommendation for collaborative activities using the characterization of learners based on learning paths (LPs). The approach emphasizes the formation of groups by applying the k-means algorithm, associated with three distance metrics of similarity (i.e., Euclidean, Manhattan, and cosine) in conjunction with the attributes derived from LPs. The framework was validated through the implementation of an M-Cluster tool. The M-Cluster presents three solution options, which can be visualized in a descriptive manner or via a bubble graph; it is the teacher who chooses the most acceptable solution for each case. The results of the case study indicate that the tool shows promise for improving the performance of students to up to 75%.