Medical training simulation system to assist novice physicians in diagnostic problem solving

  • Authors:
  • Leila Weitzel Martins;Joaquim Teixeira De Assis;André Soares Monat

  • Affiliations:
  • Energy Planning Program, Federal University of Rio de Janeiro, Technology Center, Brazil;Polytechnical Institute, State University of Rio de Janeiro, Brazil;Polytechnical Institute, State University of Rio de Janeiro, Brazil

  • Venue:
  • NN'05 Proceedings of the 6th WSEAS international conference on Neural networks
  • Year:
  • 2005

Quantified Score

Hi-index 0.00

Visualization

Abstract

Medical diagnosis can be viewed as a task that allows physicians to make predictions about features of clinical situations and to determine appropriate course of action. Medical training simulations systems provide students and novices with opportunities to practice critical tasks. The main goal of this paper is investigate the feasibility of typical technique of Pattern Recognition for Meningitis classification in order to assist health professionals in diagnostic problem. We use non-linear mapping structures, which is based on the function of human brain, which so-called Artificial Neural Network (ANN). The early meningitis diagnosis is important as the treatment differ, depending whether meningitis is caused by a virus or bacterium and because of the high severity of illness. The results showed that the use of backpropagation learning with the topology and parameters adopt was able to classify accurately all-training set. Therefore, ANN models provided a robust tool for prediction meningitis diagnostic cluster. Future work will be oriented toward to investigate of structure of the data according to their interrelation, and the feasibility to reduce the feature space dimension.