Sistema tutor inteligente baseado em aprendizado de máquina para ensino-aprendizagem de manutenção de software

• Ano de defesa: 2023
• Autor(a) principal: Francisco, Rodrigo Elias
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Uberlândia; Brasil; Programa de Pós-graduação em Ciência da Computação
• Palavras-chave em Português: Sistema Tutor Inteligente; Manutenção de Software; Aprendizado de Máquina; Intelligent Tutoring System; Software Maintenance; Machine Learning; CNPQ::CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO::METODOLOGIA E TECNICAS DA COMPUTACAO::SISTEMAS DE INFORMACAO; Computação; Sistemas tutoriais inteligentes; Aprendizado do computador; Software - Manutenção; ODS::ODS 4. Educação de qualidade - Assegurar a educação inclusiva, e equitativa e de qualidade, e promover oportunidades de aprendizagem ao longo da vida para todos.
• Link de acesso: https://repositorio.ufu.br/handle/123456789/41374; http://doi.org/10.14393/ufu.te.2024.5007

There is a market demand for qualified professionals to work with SM. The training of these professionals is quite complex, as they need to be able to carry out certain activities, such as understanding the source code and diagrams and manipulating techniques and tools aimed at SM. The SM teacher faces difficulties in offering adequate support to classes at a viable time, which makes ITS for SM an up-and-coming category of systems. The literature points out challenges on the topic and indicates low use of ML in ITS for SM. This thesis proposes an architecture for ITS focusing on SM. It addresses the use of ML in the Tutor and Student Modules and their integration, contributing to research challenges. The Student Module works with identifying types of SM students using Clustering. The Tutor Module works with the DM recommendation of SM using RL through the Q-Learning algorithm. The thesis also presents the modeling of EKM based on SM content, which contributed to evaluating the Tutor and Student Modules based on the SM knowledge dimension. The results indicate that the K-Means algorithm is suitable for the Student Module and that its integration with the Tutor Module, under certain conditions, brings high gains in the efficiency of DM recommendation. The evaluations were conducted using a data set of student SM capabilities based on a real data set of student performance and computer simulation. They showed that the ITS proposal brought significant results regarding the efficiency of DM recommendations for SM.