This study proposes a methodology, established in quantitative procedures, to assist the diagnostic of individuals with heart disease. The results obtained in this study using "Neural Networks" had been compared with the results of other authors. A percentage of average correct diagnosis of 91.0 % was reached, whereas other studies using the same database had reached until 83.5 %. Others techniques of classification of standards known in literature had also been used, called "Discriminate Analysis" and "C4.5 Algorithm", to establish comparisons with the results obtained here using "Neural Networks". The methodology of division of the sets of training/generalization suggested promoted improvements in all the three used techniques of classification of standards, with the "Neural Networks" showing the best performance.
mutual information; neural networks; heart disease
