Özet:

The influence of the number of complex components of fast Fourier transformation in analyzing the polar maps of radionuclide examination of myocardium at rest and stress on the functional efficiency of the system of diagnostics of pathologies of myocardium was explored,  and there were defined their optimum values in the information sense, which allows  increasing the efficiency of the algorithms of forming the diagnostic decision rules by reducing the capacity of the dictionary of features of recognition. The information-extreme sequential cluster algorithms of the selection of the dictionary of features, which contains both quantitative and category features were developed and the results of their work were compared. The modificatios of the algorithms of the selection of the dictionary were suggested, which allows increasing both the search speed of the optimal in the information sense dictionary and reducing its capacity by 40 %.  We managed to get the faultless by the training matrix decision rules, the accuracy of which is in the exam mode asymptotically approaches the limit. It was experimentally confirmed that the implementation of the proposed algorithm of the diagnosing system training has allowed to reduce the minimum representative volume of the training matrix from 300 to 81 vectors-implementations of the classes of recognition of the functional myocardium state. Author Biographies Anatoly Dovbysh, Sumy State University. Rimsky-Korsakov str., 2, Sumy, Ukraine, 40007 Doctor of technical sciences, Professor, Head of the department Department of Computer Science

Anahtar Kelimeler: