APPLICATION OF NEUROET NETWORK TECHNOLOGY FOR NOISE
SPECTROSCOPY OF ELECTROCARDIOGRAM
Azerbaijan State Oil and Industry University, Azadlig 20, Baku, Azerbaijan
aynur.Jabiyeva@outlook.com
Abstract
: The neural network is one of the special cases of image recognition methods, the method
of discriminant analysis, and clustering methods. Neural networks are not programmed, but they are trained. This is one of the main
advantages of neural networks over traditional algorithms. The most famous application is classification tasks. We used neural networks
to classify diseases in medicine. The most difficult problem is the definition of input parameters. With enough input, this can take a
long time to learn.
As for the literature, there are a lot of them, of which, as classical ones, it can be noted-C. Khaikin-neural networks (full
course), Callan-basic concepts of neural networks, Osovski-neural networks for information processing, etc.
Keywords
: noise, spectroscopy, parametrization, diagnostic indices, chaotic signal, autocorrelation function, power spectrum, neural
network.
Overview
As a method of nonlinear dynamics, which makes it possible to extract a prisoner in the signals produced by the
human body, the method of noise spectroscopy is considered. New possibilities of noise spectroscopy in the recognition
of 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.
Conclusions
An analysis of the experimental data shows that the modular structure of the neural network construction is the
most effective in detecting pathologies and arrhythmias of the heart. Therefore, it was chosen as the basis for constructing
a neural network block for the hardware-software complex of the EKS analysis.
The results of testing and approbation of the modular structure of INS showed a higher reliability of
cardiocomplex separation into classes "norm" and "pathology" in comparison with the device "Cardiovisor-6C" known
and used in practical public health. In particular, the modular INS provided sensitivity by 5%, and specificity - by 26%
higher than "Cardiovisor-6C".
The presentation of electrocardiographic signals in the form of successive irregularities allows us to use the
method of noise spectroscopy when analyzing such signals.
A chaotic signal, represented by a time series in the case of noise spectroscopy, makes it possible to parametrize
these signals and determine informative diagnostic indices characterizing the functional state of the cardiovascular
system. The set of informative parameters, as well as the frequency of signal sampling, which determines the dynamics
of changes in these parameters, makes it possible to classify heart diseases with the help of a neural network.
Aynur J.Jabiyeva
Do'stlaringiz bilan baham: |