Representação de conhecimento relacionado ao prognóstico de pacientes com câncer de pulmão

• Ano de defesa: 2013
• Autor(a) principal: Minelli, Leonardo
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Santa Maria; BR; Ciência da Computação; UFSM; Programa de Pós-Graduação em Informática
• Palavras-chave em Português: Mineração de dados; Representação de conhecimento; Prognóstico; Data mining; Knowledge representation; Prognosis; CNPQ::CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO
• Link de acesso: http://repositorio.ufsm.br/handle/1/5430

Due to the gradual increase of cancer cases, there is a need to use computing resources to accurately retrieve the information recorded in databases. We can highlight the importance of the retrieved information for a specialist to define the prognosis of new cancer cases. This work presents a way to represent knowledge of cancer registries with emphasis on prognosis. Collecting these records, the aim was to use methods of knowledge representation with data mining techniques to find patterns in the data stored for patients lifetime in similar situations. The work developed is focused on the generation of association rules to find the best patterns on these registries in order to measure the prognosis of patients and subsequent health expert s conclusions. This scope was given through the evaluation of other forms of knowledge representation in databases, and the results were subjected to validation with parallel evaluation with international organizations of oncology and health publications.