Mineração de dados educacionais para a gestão de cursos massivos

Resumo (EN)

In relation to traditional distance learning courses, as also MOOCs (Massive Open Online Courses), the Learning Management System (LMS) show a significant volume of data regarding navigation, participation and progress of students, enabling the carrying out of several different analyses related to the standard, or set of standards, of behaviour among this group. The analysis of this significant volume of data is not feasible without the aid of appropriate computer resources, and here it is essential that tools be made available to help the managers of such courses in their tasks of analysing and interpreting this data, and also in establishing relationships between the items of data, so that it may be possible to draw up and define strategies for action in each domain of application. The core purpose of this thesis is that of describing and analysing the contributions and also the limitations of the application of educational data mining (EDM) methods to identify new and relevant knowledge about the standard, or set of standards, of behaviour among students in massive courses. The managers no longer take decisions in a subjective and empirical way because, based on the standards that have been discovered, it would be possible to generate useful knowledge to help with decision making, seeking to reduce the dropout rate and also to improve the teaching-learning process. In methodological terms, this is an applied research study in which the main procedure used is that of research-action. The application of EDM algorithms, specifically decision trees and clustering, make it possible to discover relevant standards of behaviour for the management of massive courses, such as the kind of attribute that has made the most significant contribution towards the students’ passes, and also the patterns of behaviour for groups of students that have failed, elements that can be very useful indeed, so that the tutorship team may address, with greater attention, those very students that most need help throughout the course.