Um modelo computacional de apoio à análise da opinião de alunos sobre práticas docentes por meio da mineração de dados educacionais

• Ano de defesa: 2017
• Autor(a) principal: Santos, Fábio de Paula
• Orientador(a): Silveira, Ismar Frango
• Banca de defesa: Silva, Leandro Augusto da , Omar, Nizam , Araújo Junior, Carlos Fernando de , Yamamoto, Cláudio Haruo
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Presbiteriana Mackenzie
• Programa de Pós-Graduação: Engenharia Elétrica
• Departamento: Escola de Engenharia Mackenzie (EE)
• País: Brasil
• Palavras-chave em Português: mineração de dados educacionais; aprendizado de máquina; análise de sentimentos; avaliação institucional docente
• Área do conhecimento CNPq: CNPQ::CIENCIAS EXATAS E DA TERRA; CNPQ::LINGUISTICA, LETRAS E ARTES::LINGUISTICA; CNPQ::CIENCIAS HUMANAS::EDUCACAO::PLANEJAMENTO E AVALIACAO EDUCACIONAL
• Link de acesso: http://dspace.mackenzie.br/handle/10899/26515

The Institutional Teacher’s Evaluation besides being a legal need in higher education is an important moment for any Educational Institution. Traditionally, questionnaires with closed answer questions are used for this purpose and many times the evaluation is left to a secondary place. This work proposes a computational model based in machine learning techniques and Sentiment Analysis that allows increasing the scope of this evaluation when allowing the use of open and textual questions. The application of these techniques in Educational Data Mining context provides basis to decision-making based on the students’ opinions. For this purpose, as proof of concept, a mining of a student’s opinion survey from a Vocational High School in Brazil was held and categorized their sentiments as positive or negative in relation to their lecturers’ techniques with supervised machine learning approach. This model also contemplates clustering analysis to find categories of analysis of student opinions through an unsupervised Learning Machine model. As a conclusion it was proven that the use of tools for textual analysis of open questions is possible and it to speeds up the decision-making of institutional evaluations.