ARTIFICIAL NEURAL NETWORKS UTILIZATION FOR THE STUDY OF ACUTE MYOCARDIAL INFARCTION PROPENSITY.

Abstract

Introduction: The advances in technology have allowed the modern society a relative comfort life. This lifestyle has contributed to the appearence of many factors that can lead to body dysfunctions, bringing serious health complications, especially including Acute Myocardial Infarction (AMI). Studies on Artificial Intelligence has been emphasized (stressed) in the development of methods and solutions to complex problems that could not be solved by traditional programming or by simple clinical history taking.  Methodology: An assessment tool on the risk factors for AMI was developed and 296 questionnaires were applied to individuals of both sexes, pertaining to one of two groups: subjects hospitalized in Surgical Unit B, Hospital Santa Clara and subjects who work in the Administrative Center of the Irmandade Santa Casa of Porto Alegre. These collected data were stored in a software applicative and used later to carry out the Artificial Neural Networks (ANN) with 14 neurons in the entry and 1 neuron in the output layer. Results: The results of this study suggests that the ANN MultiLayer Perceptron trained and validated can be used to aid the recognition of the inter inter-relationship of risk factors, which may be used in prevention, diagnosis and support to the identification of AMI in a new set of individuals. Conclusions: The ANN have proved a valuable tool for the recognition and support in studies related to AMI.