Um sistema multiagente de identificação e caracterização de relações sociais de alunos em um ambiente virtual de aprendizagem

Abstract

The increase use of Learning Management Systems (LMS) as disseminators of knowledge aids students in developing knowledge and provides an interaction between people through computational tools mediated by the teacher (mediator or tutor). However, many LMSs does not have a mechanism or tool to aid the teacher to get detailed information about the social interaction between students and their group. This complicates the teacher’s awareness with regard to social relations established within the discipline. Based on this information, this paper proposes and details an approach to this problem that makes use of a Multi-Agent System architecture directed to the analysis of the interactions of groups. As proof of concept, a prototype was developed as using a software methodology for Multi-Agent Systems in order to analyze the interactions between students that occurred in the forums and messages of a course in the LMS Moodle and present this information to the teacher in the form of sociogram or table. In this way, the teacher can visually monitor the interactions of students within the course, and adopt tactics that promote social inclusion. To validate the work a case study was performed, a class of a course of Distance Education in a Higher Education Institution (HEI). According to the results obtained in simulated tests and a real class, it is clear that the graphic visualization of interactions of students in the forums and messages, when combined with the proactivity of agents, can corroborate the work done by mediators. The vizualization of the interactions provides to the teachers a more comprehensive, clear and quick view of social interactions established in the discipline, facilitating a option to intervene intentionally through teaching strategies such as to encourage students to develop the communication, interaction and the creation of new groups.