TY - JOUR AU - Skrupsky, S. Yu. PY - 2016/08/25 Y2 - 2024/03/28 TI - EXPERIMENTAL INVESTIGATION OF METHOD FOR THE SYNTHESIS OF NEURO-FUZZY MODELS IN A PARALLEL COMPUTER SYSTEM JF - Radio Electronics, Computer Science, Control JA - RIC VL - 0 IS - 2 SE - Neuroinformatics and intelligent systems DO - 10.15588/1607-3274-2016-2-7 UR - http://ric.zntu.edu.ua/article/view/76209 SP - AB - The article deals with the problem of the development of the non-linear model describing dependences between the characteristics of a<br />system, in which synthesis of neuro-fuzzy networks is realized, the parameters of the investigated method and the time spent on execution of<br />the models synthesis. The object of research is a synthesis of neuro-fuzzy models for individual prediction of the hypertensive patient state.<br />The subject of research is a parallel computer system that performs the method of neuro-fuzzy networks synthesis. The purpose of the work<br />is to improve the efficiency of parallel computer systems solving the problems of medical direction. A non-linear model to predict the time<br />used by a parallel system to perform the method of neuro-fuzzy network synthesis and thus to execute a rational choice of the computer system<br />resources has been proposed. The software that implements the proposed model has been developed. Experiments confirming the adequacy of the proposed model have been executed. The experimental results allow us to recommend the application of the developed model in practice. ER -