Ontology-based group assessment analytics framework for performances prediction in project-based collaborative learning

Abstract

Abstract This article introduces an ontology-based framework for group assessment analytics that investigates the impact of intra-group interactions on group performance within the context of project-based collaborative learning (PBCL). Additionally, it aims to predict learners’ performance based on these interactions. The study involved 312 first-degree students specializing in transportation and technology engineering. The framework collects interaction data from discussion forums and chat rooms, conducts comprehensive data analysis, and constructs prediction models using supervised learning methods. The results unequivocally demonstrate that intra-group interactions significantly affect group performance in PBCL. The prediction model, with an accuracy metric of 0.92 and a final test score of 0.77, supports the credibility of the findings. Notably, the framework utilizes an ePortfolio specifically designed for group assessments, effectively managing both assessment and group data. This framework provides educators with a robust tool to assess group performance, identify areas requiring improvement, and contribute to shaping informed student learning outcomes. Furthermore, it empowers students by enabling them to receive feedback on their collaborative efforts, fostering enhanced interaction skills. These findings carry significant implications for the development and implementation of PBCL environments, offering educators valuable insights for evaluating student progress and making strategic decisions.

Keywords

• Collaborative learning
• Project-based collaborative learning (PBCL)
• Group assessment
• Analytics framework
• Intra-group interactions
• Ontology-based