Uma Análise Exploratória da Influência dos Projetos Pedagógicos dos Cursos Superiores no Resultado do Enade por meio de Mineração de Textos e Aprendizado de Máquina

• Ano de defesa: 2022
• Autor(a) principal: Charles Andre Profilio dos Santos
• Orientador(a): Liana Dessandre Duenha Garanhani
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Fundação Universidade Federal de Mato Grosso do Sul
• País: Brasil
• Palavras-chave em Português: Aprendizado de Máquina, Mineração de texto, Educação Superior
• Link de acesso: https://repositorio.ufms.br/handle/123456789/5071

A higher education course is guided by the Pedagogical Project of the Course (PPC), which suggests the training expected for the graduates, both in the professional and humanistic aspects, according to the current national curriculum guidelines. To evaluate undergraduate courses and higher education institutions, the Ministry of Education (MEC) uses some quality indicators, such as the National Student Performance Exam (Enade), with an assessment applied every three years to students graduating from each course, which aims to assess the quality of undergraduate education in the country by assigning a concept to each evaluated course. This concept and the other evaluation reports resulting from Enade help the managers of higher education institutions, course coordinators and professors to act to improve their pedagogical projects, physical infrastructure, human resources and other aspects that impact student training. This work proposes an analysis of the pedagogical projects of courses using machine learning, to assist in the understanding of how its content impacts the evaluation, more specifically, the concepts Enade Track and Enade Continuous. The analysis was applied to PPC of Computer Science and Information Systems courses, however the the methodology applies to other courses, by replicating the method on new training data. The experimental results showed that it is possible to predict the Enade Range Concept with an accuracy of ≈ 80%, and the Enade Continuous concept with an average absolute percentage error of ≈ 11%.