MONITUM: um sistema proativo para monitoramento e avaliação das atividades de tutoria a distância em AVAs

Abstract

The use of Virtual Learning Environments (VLEs) in Distance Education (DE) has generated an increasing volume of data coming from interactions between the actors of this process. The distance tutor is the actor responsible for mediating the students’ learning process and for promoting interaction in the VLEs. The volume of data generated from these interactions, if properly exploited, can provide an understanding of the relationship of influence between the performance of distance tutors and the effective participation of students in VLEs. In this context, the research objective of this thesis was defined as the development of a proactive system for monitoring and evaluation of distance learning activities, through behavioral indicators in VLEs. The method used in this research was based on the six-dimensional Learning Analytics methodology, which defined stakeholders, objectives, data, instruments, external constraints and internal limitations. The analyzes were carried out using historical data from a technological course and ten undergraduate courses offered by the Secretariat for Distance Education of the Federal University of Rio Grande do Norte. In all, three samples were used. The first sample consisted of information from 38 tutors the distance and the second sample consisted of information from 2,227 students, being that the data from the two samples belonging to 62 classes. The third sample consisted of information from 353 tutors at a distance, belonging to 1,281 classes. As main results of this thesis, we have: (i) construction of a knowledge model (ontology) capable of mapping the existing relationships between the behavioral attributes of the distance tutors and the effective participation of the students, through Pearson’s Correlation Coefficient ; (Ii) construction of a Web Service capable of processing the information related to the attributes of the distance tutors, classifying their performances through the k-Means and Farthest First techniques and making recommendations based on the knowledge model; and (iii) creation of a plugin for Moode that allows the visualization of the data processed by the Web Service and that serves as a subsidy for the managers of said courses.