Análise do impacto da evasão e retenção no ensino superior utilizando cadeias de Markov absorventes

• Ano de defesa: 2022
• Autor(a) principal: SANTOS, Juliana Ferreira dos
• Orientador(a): ALBUQUERQUE JÚNIOR, Gabriel Alves de
• Banca de defesa: ANDRADE, Ermeson Carneiro de, ARAUJO, Jean Carlos Teixeira de
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal Rural de Pernambuco
• Programa de Pós-Graduação: Programa de Pós-Graduação em Informática Aplicada
• Departamento: Departamento de Estatística e Informática
• País: Brasil
• Palavras-chave em Português: Análise de sobrevivência; Cadeia de Markov; Ensino superior
• Área do conhecimento CNPq: CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO
• Link de acesso: http://www.tede2.ufrpe.br:8080/tede2/handle/tede2/8676

Dropout and retention are recurring problems in undergraduate courses, capable of causing idle resources or generating revenue loss in universities. Survival analysis allows verifying the influence of events, such as dropout, completion and student bond, according to the time and probability of occurrence. It allows finding solutions to solve these problems in advance and avoid loss to universities. This work aims to analyze the dropout, completion and detachment of undergraduate students, through the proposal of a Markov Chain model to perform the survival analysis of students throughout undergraduate courses. The absorbing Markov Chain proposed in this work simulates academic progress, through states that represent the semesters in which students are bound to the course, with the inclusion of retained states. In the case studies, with graduation data from a Brazilian public university, the analyses identified differences in the behavior of dropout for courses in the areas of agrarian, computing and health. In addition, it was analyzed how the semester in which the student was retained impacts the probabilities of dropout, completion and detachment. In the analyzes considering the categories course, gender and race, it was identified how the difference in the behavior of students in these categories influence dropout, completion and detachment. Experiments were also carried out that showed that to decrease the overall dropout rate, controlling retention in the first semesters has a greater impact than controlling dropout itself.