EduColab: Sistema de recomendação educacional para promover a colaboração em fóruns de discussão

Abstract

A software can recommend interactions among students in a Virtual Learning Environment (VLE) for the purpose of supporting collaborative learning. However, assessing collaboration in such environment to continuously encourage learning remains a challenge. This thesis describes a research which addresses the development of an Educational Recommender System (ERS) that adopts the Conversational Analysis (CA) in order to identify students’ collaboration and makes recommendations for promoting collaborative learning in discussion forums of a VLE. Based on experiments, a CA strategy is defined to identify the level of collaboration among students, possibilities for intervention in favor of collaborative learning are pointed out, the unsupervised machine learning in indexes resulting from CA is applied to validate and refine the pre-established recommendation strategy, and the results achieved are presented, including the EduColab, which is the ERS developed aiming to promote collaboration in a VLE discussion forums.