Neural Network System for Medical Data Approximation
Andrey N. Astafyev, Sergey I. Gerashchenko, Marina V. Markuleva, Геращенко Михаил Сергеевич · 2020
The paper considers the technique of medical data using neural network approximation. The method of neural networks training uses the Kachmazh's algorithm to solve linear equations systems, justifies the choice of the algorithm for solving linear equations systems using the iterative Kachmazh's method. A review of solving linear equations systems method using the Kachmazh's algorithm is carried out, the conditions of convergence are considered. The analysis of the algorithm application method in neural networks is made. An overview of the system structural scheme for medical data approximation based on the considered technique is given. The necessity of neural network systems application for medical diagnostics is substantiated.