Aplicação de técnicas de inteligência artificial na formação de equipes de aprendizagem colaborativa baseada em perfil de estilos de aprendizagem e papéis de equipe

Abstract

The inherently social nature of human beings leads us to seek ways to interact with our fellow human beings. The emergence of virtual communication channels (such as social networks and instant messaging systems) and the popularization of the internet and mobile devices have incorporated this connectivity to practically all of us, especially children and teenagers, who were born in this environment. Today, being connected is a necessity, including in education. This need to be connected and get good results makes it important to identify factors that affect the joint performance of students and find efficient ways to organize study teams. This dissertation investigates whether collaborative study teams formed based on individual profiles that match students' learning styles and team roles perform better than casually formed study teams. Individual student profiles are identified from their responses to Felder's learning styles questionnaires and Belbin's team roles. A genetic algorithm uses these profiles to generate balanced teams according to optimization rules. To assess performance, teams perform collaborative activities and answer a final satisfaction questionnaire related to their perception of teammates and the activity. Compared to randomly formed teams, the teams generated by the AG based on the profile composed of learning styles and team roles have better academic performance and greater individual satisfaction.