Forecasting of COVID-19 Epidemic Process by Support Vector Machine Method in Ukraine and Neighboring Countries

Abstract

The global pandemic has affected all areas of life. Scientifically based management decisions to reduce epidemic morbidity not only increase their efficiency, but also save costs aimed at eliminating the virus. For this, mathematical modeling of epidemic processes is used. The most accurate approach to predicting incidence is machine learning. To study and predict the dynamics of the infectious morbidity of COVID-19, a regression model was built based on the support vector machine. The following countries were selected to verify and check the adequacy of the model: Belarus, Hungary, Moldova, Poland, Romania, Russia, Slovakia and Ukraine. Forecasting in these countries allows us to study the impact of the epidemic of neighboring countries on the dynamics in Ukraine, as well as to determine the accuracy of the developed model.