Conventional Neural Network for Models of Medical Risk Classifiers with Synergy Channels

The purpose of research is to improve the quality of predicting cardiovascular risks by using synergistic channels formed by convolutional neural networks.

Methods. To predict the functional state of living systems, it is proposed to supplement the risk factors for cardiovascular diseases with virtual factors, taking into account the influence of real risk factors on each other. A method has been developed for constructing a convolutional neural network intended for the analysis and classification of images constructed from the results of modeling synergetic channels, which is characterized by the use of the minimum dimension of convolutional filters and methods for completing the definition of the original images.

Results. To test the proposed method and model of the classifier, an experimental group was formed of patients with three gradations of the risk of coronary heart disease (IHD) and a classifier of IHD risk was built on the basis of a convolutional neural network in three gradations. As a concomitant disease that can stimulate a synergistic effect, this model uses vibration disease, and the electromagnetic field is taken as an environmental factor contributing to the synergistic effect. Traditional coronary heart disease risk factors and descriptors determined on the basis of electrocardiographic studies were selected as two more segments of risk factors. Considering that locomotive crew drivers were selected as the object of the study, risk factors associated with their professional activities were taken as the fifth segment of risk factors leading to the onset and development of coronary heart disease.

Conclusion. In the course of experimental assessment and as a result of mathematical modeling, it was shown that when all risk factors were used, confidence in the correct prognosis of IHD exceeds 0.8 for all observation groups and for all indicators of classification quality. Indicators of forecasting quality are higher than those of the well-known SCORE forecasting system, on average, by 14%.

1. Smirnova M. D., Fofanova T. V., Ageev F. T., eds. Prognosticheskie faktory razvitiya serdechno-sosudistyh oslozhnenij vo vremya anomal'noj zhary 2010 g. (kagornoe nablyudatel'noe issledovanie) [Predictive factors for the development of cardiovascular complications during the heat wave of 2010 (Cahors observational study)]. Kardiologicheskij vestnik = Cardiological Bulletin, 2016, no. 11(1), pp. 43-51.

2. Efremov M. A. , Filist S. A., Shatalova O. V., eds. Gibridnye nechetkie modeli dlya prognozirovaniya vozniknoveniya i oslozhnenij arterial'noj gipertenzii s uchetom energeticheskih harakteristik bioaktivnyh tochek [Hybrid fuzzy models for predicting the occurrence and complications of arterial hypertension, taking into account the energy characteristics of bioactive points]. Izvestiya Yugo-Zapadnogo gosudarstvennogo universiteta. Seriya: Upravlenie, vychislitel'naya tekhnika, informatika. Medicinskoe priborostroenie = Proceedings of the Southwest State University. Series Control, Computer Engineering, Information Science. Medical Instruments Engineering, 2018, vol. 8, no. 4 (29), pp. 104-119.

3. Filist S. A., Kurochkin A. G., Zhilin V. V., eds. Ispol'zovanie gibridnyh nejrosetevyh modelej dlya mnogoagentnyh sistem klassifikacii v geterogennom prostranstve informativnyh priznakov [Using hybrid neural network models for multi-agent classification systems in a heterogeneous space of informative features]. Prikaspijskij zhurnal: upravlenie i vysokie tekhnologii. Nauchno-tekhnicheskij zhurnal = Caspian Journal: Management and High Technologies. Scientific and Technical Journal, 2015, no. 3 (31), pp. 85-95.

4. Filist S. A., Tomakova R. A., Zar Do Yaa. Universal'nye setevye modeli dlya zadach klassifikacii biomedicinskih dannyh [Universal network models for biomedical data classification problems]. Izvestiya Yugo-Zapadnogo gosudarstvennogo universiteta = Proceedings of the Southwest State University, 2012, no 4 (43), pt. 2, pp. 44-50.

5. Filist S. A., Shatalova O. V., Efremov M. A. Gibridnaya nejronnaya set' s makrosloyami dlya medicinskih prilozhenij [Hybrid neural network with macro layers for medical applications]. Nejrokomp'yutery. Razrabotka i primenenie = Neurocomputers. Development and Application, 2014, no. 6, pp. 35-39.

6. Kiselev A. V., Petrova T. V., Degtyarev S. V., eds. Nejrosetevye moduli s virtual'nymi potokami dlya klassifikacii i prognozirovaniya funkcional'nogo sostoyaniya slozhnyh system [Hybrid Deciding Modules with Virtual Streams for Classification and Prediction of Functional State of Complex Systems]. Izvestiya Yugo-Zapadnogo gosudarstvennogo universiteta = Proceedings of the Southwest State University, 2018, vol. 22, no. 4, pp. 123134.

7. Komlev I. A., Shatalova O. V., Degtyarev S. V., Serebrovsky A. V. Prognozirovanie i otsenka stepeni tyazhesti ishemii serdtsa na osnove gibridnykh nechetkikh modelei [Prediction and assessment of the severity of heart ischemia on the basis of hybrid fuzzy models]. Izvestiya Yugo-Zapadnogo gosudarstvennogo universiteta. Seriya: Upravlenie, vychislitel'naya tekhnika, informatika. Medicinskoe priborostroenie = Proceedings of the Southwest State University. Series: Control, Computer Engineering, Information Science. Medical Instruments Engineering, 2019, vol. 9, no. 1 (30), рр. 133-145.

8. Filist S. A., Shutkin A. N., Shkatova E. S., Degtyarev S. V., Savinov D. Yu. Model' formirovaniya funkcional'nyh sistem s uchetom menedzhmenta adaptacionnogo potenciala [Model of the formation of functional systems taking into account the management of adaptive potential]. Biotekhnosfera = Biotechnosphere, 2018, no. 1 (55), pp. 32-37.

9. Aksenov S. V., Kostin K. A., Ivanova A. V., Liang J., Zamyatin A. V. Diagnostika patologij po dannym videoendoskopii s ispol'zovaniem ansamblya svertochnyh nejronnyh setej [Diagnosis of pathologies according to video endoscopy using an ensemble of convolutional neural networks]. Sovremennye tekhnologii v medicine = Modern technologies in medicine, 2018, no. 10(2), pp. 7-19. https://doi.org/10.17691/stm2018.10.2.01.

10. Nibali A., He Z., Wollersheim D. Pulmonary nodule classification with deep residual networks. Int J Comput Assist Radiol Surg, 2017, no. 12(10), pp. 1799-1808. https:// doi.org/10.1007/s11548-017-1605-6

11. Tajbakhsh N., Gurudu S. R., Liang J. Automatic polyp detection in colonoscopy videos using an ensemble of convolutional neural networks. IEEE 12th International Symposium on Biomedical Imaging (ISBI), 2015. https://doi.org/10.1109/isbi.2015.7163821

12. LeCun Y., Kavukcuoglu K., Farabet C. Convolutional networks and applications in vision. Proceedings of 2010 IEEE International Symposium on Circuits and Systems, 2010. https://doi.org/10.1109/iscas.2010.5537907

13. Sozykin A. V. Obzor metodov obucheniya glubokih nejronnyh setej [Overview of methods for training deep neural networks]. Vestnik Yuzhno-Ural'skogo gosudarstvennogo universiteta. Seriya: Vychislitel'naya matematika i informatika = Bulletin of the South Ural State University. Series: Computational Mathematics and Informatics, 2017, vol. 6, no. 3, pp. 28-59. https://doi.org/10.14529/cmse170303

14. Bredihin A. I. Algoritmy obucheniya svertochnyh nejronnyh setej [Convolutional neural network learning algorithms]. Vestnik Yugorskogo gosudarstvennogo universiteta = Bulletin of the Ugra State University, 2019, vol. 1 (52), pp. 41-54.

15. Kobyakova O. S., Deev I. A., Kulikov E. S., Starovojtova E. A., Malyh R. D., Balaganskaya M. A., Zagromova T. A. Hronicheskie neinfekcionnye zabolevaniya: effekty sochetannogo vliyaniya faktorov riska [Chronic noncommunicable diseases: effects of the combined influence of risk factors]. Profilakticheskaya medicina = Preventive Medicine, 2019, no. 22(2), pp. 45-50.

16. Li K., Husing A., Kaaks R. Lifestyle risk factors and residual life expectancy at age 40: a German cohort study. BMC Medicine, 2014, no. 12(1). https://doi.org/10.1186/1741-7015-12-59

17. Belyj O. V., Barinova L. D., Abramov A. M. Problemy elektromagnitnoj bezopasnosti na transporte [Problems of electromagnetic safety in transport]. Transport Rossijskoj Federacii = Transport of the Russian Federation, 2018, no. 2 (75), pp.71-73.

18. Kiselev A. V., Shatalova O. V., Protasova Z. U., Filist S. A., Stadnichenko N. S. Modeli latentnyh prediktorov v intellektual'nyh sistemah prognozirovaniya sostoyaniya zhivyh system [Latent predictor models in intelligent systems for predicting the state of living systems]. Izvestiya Yugo-Zapadnogo gosudarstvennogo universiteta. Seriya: Upravlenie, vychislitel'naya tekhnika, informatika. Medicinskoe priborostroenie = Proceedings of the Southwest State University. Series: Control, Computer Engineering, Information Science. Medical Instruments Engineering, 2020, no. 10(1), pp. 114-133. (In Russ.)

19. Dudchenko A., Ganzinger M., Kopanitsa G. Machine Learning Algorithms in Cardiology Domain: A Systematic Review. The Open Bioinformatics Journal, 2020, no. 13, pp. 25-40. https://doi.org/10.2174/1875036202013010025.