Predictive model of COVID-19 epidemic process based on neural network

Abstract

The COVID-19 pandemic, which has been going on for almost three years, has shown that public health systems are not ready for such a challenge. Measures taken by governments in the healthcare sector in the context of a sharp increase in the pressure on it include containment of the transmission and spread of the virus, providing sufficient space for medical care, ensuring the availability of testing facilities and medical care, and mobilizing and retraining medical personnel. The pandemic has changed government and business processes, digitalizing the economy and healthcare. Global challenges have stimulated data-driven medicine research. Forecasting the epidemic process of infectious processes would make it possible to assess the scale of the impending pandemic to plan the necessary control measures. The study builds a model of the COVID-19 epidemic process to predict its dynamics based on neural networks. The target of the research is the infectious diseases epidemic process in the example of COVID-19. The research subjects are the methods and models of epidemic process simulation based on neural networks. As a result of this research, a simulation model of COVID-19 epidemic process based on a neural network was built. The model showed high accuracy: from 93.11% to 93.96% for Germany, from 95.53% to 95.54% for Japan, from 97.49% to 98.43% for South Korea, from 93.34% up to 94.18% for Ukraine, depending on the forecasting period. The assessment of absolute errors confirms that the model can be used in healthcare practice to develop control measures to contain the COVID-19 pandemic. The respective contribution of this research is two-fold. Firstly, the development of models based on the neural network approach will allow estimate the accuracy of such methods applied to the simulation of the COVID-19 epidemic process. Secondly, an investigation of the experimental study with a developed model applied to data from four countries will contribute to empirical evaluation of the effectiveness of its application not only to COVID-19 but also to other infectious diseases simulations. Conclusions. The research’s significance lies in the fact that automated decision support systems for epidemiologists and other public health workers can improve the efficiency of making anti-epidemic decisions. This study is especially relevant in the context of the escalation of the Russian war in Ukraine when the healthcare system's resources are limited.