Radiotekhnika
Publishing house Radiotekhnika

"Publishing house Radiotekhnika":
scientific and technical literature.
Books and journals of publishing houses: IPRZHR, RS-PRESS, SCIENCE-PRESS


Тел.: +7 (495) 625-9241

 

Prognostication of Effect From Non-Invasive Procedure Eliminating Concretions using Hybrid Technologies of Fuzzy Logic and Neural Networks

Keywords:

V.V. Zhilin, Ali Abdo Mokhammed Avad, S.A Filist


This paper presents the results of the using of hybrid technology based on fuzzy logic and neural networks to predict the effect of therapeutic procedures of distance shock-wave lithotripsy in the treatment of urolithiasis. To perform the prognostication we have analysed characteristics of the etiology and pathogenesis of this disease, and got a set of possible outcomes of medical therapy and a list of informative features that influence the result of forecasting. A method of structuring space of signs by splitting it into several disjoint groups is developed. Assessing the impact of each of the signs for the prediction of the outcome of therapy is performed by compiling verbal expert opinions on each of the signs. After that we are performing the formalization of the knowledge gained with the developed method for constructing fuzzy membership functions. This method is based on the union of two fuzzy functions and utilization of the weighting factor to limit the maximum value of the final membership function. There are an example of formalization of expertise based on analysis of the influence of patient age for the effect of the procedure extracorporeal lithotripsy in this article, also demonstrated the form and parameters of the final functions. To implement the predictive system we have developed an program package in the Matlab, a modular structure and purpose of each of the five interconnected modules shown in this paper. By using this software package im-plemented forecasting the results of extracorporeal lithotripsy in patients in the control sample, and comparation performance of the our system with alternative ways to predict, based on the technologies of fuzzy inference and learning of neural networks. Using a hybrid forecasting system we got an diagnostic efficacy 0.91, showing the result of much higher than the other two models.
References:
June 24, 2020
May 29, 2020

© Издательство «РАДИОТЕХНИКА», 2004-2017            Тел.: (495) 625-9241                   Designed by [SWAP]Studio