Churyumova I. Diagnostic of Borderline Cases of Human Organism by Methods of Computational Intelligence

The thesis is directed on solution of the actual scientific and practical task - rising of probability of straightening setting of the diagnosis. The dissertation is devoted to development of a method of diagnostic of borderline cases of a human organism. In a thesis choice of the way of dimensionality reduction medical data sets is proved in case of deficiency of the information. The adaptive method of fuzzy clusterization for multiprocessing and serial processing of medicobiological data sets of a human organism is developed. Sort of the metrics is updated at clusterization of the medical data. The frame of the diagnostic system of objective formulation of diagnosis is synthesized. Feature of this system is in impossibility of influence of the doctor on the process of diagnostic. System operation is experimentally tested on various under the profile of borderline cases contingents of patients.