Using Agents for Training Students Collaborative Skills
Rosanna Costaguta, Patricio García Báez, Analı́a Amandi · IEEE Latin America Transactions · 2011
Although the advantages of Computer Supported Collaborative Learning systems are widely recognized, it is also known that using these systems does not guarantee that collaborative learning occurs. The collaborative skills that the students in a group are able to show determine, among other factors, the success or failure of the learning experience. This paper presents a multi-agent model developed to recognize conflicts in a group dynamics, and personal training skills of collaboration expressed by the students. The Interaction Process Analysis method is applied in the detection of conflicts in group dynamics. Then, Bayesian networks modeling the collaboration features of each student are used to determine an appropriate training strategy for each individual. The metrics applied after the experiments show a high degree of efficiency in the model created.