Modelo GAC-BRASIL: algoritmo genético aplicado à similaridade para raciocínio baseado em casos: estudo de caso em diagnóstico de cardiopatia isquêmica
AUTOR(ES)
Sandro Moretti Correia de Almeida
DATA DE PUBLICAÇÃO
2006
RESUMO
The goal of Case-Based Reasoning (CBR) is to reuse known solutions to solve new problems by searching for similar cases in a base of cases. Similarity establishing criteria is one of the major issues of this Artificial Intelligence technique. The CBR module of Artificial Intelligence Applied in the Modeling and Implementation of a Virtual Medical Office Project (IACVIRTUAL) uses a global similarity function that computes the weighed sum of the local similarity functions previous results for each symptom or characteristic between the cases. The weighs that indicated the relevance used by this function have been indicated by an Ischemic Cardiopathy expert. Although this method is largely applied, this definition totally relies on the skill and medical knowledge of one or more experts, therefore is subjective and non statistics oriented. Genetic Algorithm is a Learning Machine method based on Evolutionary Computation that covers the search space by combining the best fitness structures survival and the information exchange between them, in a structured and random way as well. Such search mechanism emulates the natural genetic and the natural selection. This work defines a model of Genetic Algorithm applied to Case-Based ReAsoning SImiLarity named GAC-BRASIL, and it will present a case study validated by a Cardiology medical expert. In this validation with real Ischemic Cardiopathy cases, the GAC-BRASIL model has been used to create a system to find the most appropriated weighs for each symptom or characteristic to be used on the global similarity function of the CBR module of IACVIRTUAL Project.
ASSUNTO(S)
ischemic cardiopathy genetic algorithm algoritmos genéticos; raciocínio; aprendizado do computador; diagnóstico; cardiopatia diagnóstico médico raciocínio baseado em casos cardiopatia isquêmica ciencia da computacao aprendizagem de máquina cálculo de similaridade similarity calculation medical diagnosis learning machine case-based reasoning
ACESSO AO ARTIGO
http://www.bdtd.ucb.br/tede/tde_busca/arquivo.php?codArquivo=473Documentos Relacionados
- Raciocínio baseado em casos: conceitos e aplicações
- Raciocínio baseado em casos: uma abordagem utilizando o sistema imune artificial
- Raciocinio baseado em modelo aplicado ao diagnostico de falha e performance em redes locais
- Raciocínio baseado em casos aplicado a diversos domínios de problema
- Ferramenta para desenvolvimento de sistemas de raciocínio baseado em casos