Group and students profiles to support collaborative learning in a multiagent model

Abstract

Computer Supported Collaborative Learning (CSCL) systems have recognized advantages. However, using these systems does not guarantee an effective collaborative learning. Success or failure of the learning experience depends on the collaborative skills the students show in the group. This work presents a multiagent model applied to CSCL environment, which aims both at recognizing conflicts occurring in group dynamics and also at providing personalized training of collaborative skills demonstrated by group members. Conflicts are recognized with the aid of information stored in a group model and by applying the Interaction Process Analysis method. Personalization is achieved through Bayesian networks that consider students' collaborative characteristics, stored in a collaborative student model, to elucidate the most suitable training strategy. © 2012 Springer-Verlag.