Identificação de alunos com tendência à evasão nos cursos de graduação a distância por meio de mineração de dados educacionais

• Ano de defesa: 2018
• Autor(a) principal: Schmitt, Jáder Adiél
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Santa Maria; Brasil; Educação; UFSM; Programa de Pós-Graduação em Tecnologias Educacionais em Rede; Centro de Educação
• Palavras-chave em Português: Mineração de dados educacionais; Previsão de evasão na EaD; Gestão Educacional na EaD; Educational data mining; Prediction of evasion in distance education; Educational management in distance education; CNPQ::CIENCIAS HUMANAS::EDUCACAO
• Link de acesso: http://repositorio.ufsm.br/handle/1/14229

The evasion is one of the great constant challenges in the educational context in all the modalities. On the other hand, the increase in the amount of data generated by systems such as Virtual Environment for Teaching Learning (AVEA in Portuguese) has been highlighting the area of Educational Data Mining (EDM). This fact can be verified through several studies that have been developed in this area with the purpose of predicting students with tendency to evasion. However, the vast majority detain on the technical aspects pertaining to mining. In this perspective, the work aims provide to educational manager’s strategic data, through the application of EDM, so that they can evaluate, reflect and generate actions to mitigate the evasion process. In order to do so, a bibliographical research was carried out with the intention of comprehending in detail themes related to distance education, evasion as well as its causes. Soon afterwards a Systematic Literature Review (SLR) was carried out aiming to know the technological approaches applied in the prediction of evasion. In the sequence, the aspects related to the discovery of knowledge in databases, related works and methodological aspects were approached. The development was carried out through two experiments covering three undergraduate courses in which data were used of student interactions in AVEA and data from the institution's academic management system. As results, good indexes of correct answers were obtained, thus allowing the conclusion that the proposed approach is feasible for the detection of students with a tendency to evasion in undergraduate courses. Lastly, an application for mobile devices was developed for the availability of mining data. In the assessment of the application, along with the coordinators of the evaluated courses, it was found that the data discovered are of paramount importance for management because it facilitates the follow-up of students with a tendency to evasion to carry out interventions with them in order to avoid their dropping out. The dissertation presented is part of the research line Development of Educational Softwares, of the Post-Graduate Program in Educational Technologies in Network, and generated as products the text presented here, as well as the EDM strategy created and the application developed.