ABSTRACTS OF ARTICLES OF THE JOURNAL "INFORMATION TECHNOLOGIES".
No. 6. Vol. 24. 2018

DOI: 10.17587/it.24.402-405

N. T. Abdullaev, a.namik46@mail.ru, Azerbaijan Technical University, Baku, AZ 1148, Azerbaijan Republic, Ì. Ì. Qasankuliyeva, metahasanquliyeva@rambler.ru, A. Dj. Dzhabieva, aynur.jabiyeva@outlook.com Azerbaijan State University of Oil and Industry, Baku city, az1010, Azerbaijan Republic

Application of Neuroet Network Technology for Flikker-Noise Spectroscopy of Electrocardiogram

The dynamics of many physiological processes occurring in the human body is chaotic and can be described from the positions of the theory of nonlinear deterministic systems. The randomness of the behavior of the heart rhythm, as a rule, is associated with the activity of the parametric nervous system. In the field of cardiovascular research, the analysis methods, mathematically applied to non-stationary signals, whose statistical properties change with time, are mainly isolated. Often they consist of short-time high-frequency components, accompanied by long low-frequency components. As a method of nonlinear dynamics, which makes it possible to extract the information contained in the signals produced by the human body, the method of flicker-noise spectroscopy is considered. New features of flicker-noise spectroscopy in the recognition specific features of biomedical signals are due to the introduction of information parameters. These parameters, which characterize the components of the signals under study at different frequency ranges, are necessary for the calculation of diagnostic indices. Automation of the process of diagnosing the functional state of the cardiovascular system is proposed to be realized with the help of artificial neural networks.Based on the computational experiment, dependencies were obtained for the normal state of the cardiovascular system and a number of "catastrophic" arrhythmias (ventricular tachycardia, atrial fibrillation, atrial arrhythmia). At the same time, experimental data were used from the public website www.PhysioNet.org/
For the computational experiment, a perceptron three-layer network with direct links. To learn the neural network, the Back Propagation algorithm was applied. The training time was about 240 s, the maximum network error was of the order of 0.05, the degree of training was about 0.01. To recognize the pathologies of the cardiovascular system, a modular version of the neural network building structure can be used.
Keywords: flicker noise, spectroscopy, parametrization, diagnostic indices, chaotic signal, autocorrelation function, power spectrum, neural network

P. 402–405

To the contents