ANALYSIS AND FORECASTING THE LEVEL OF CORONAVIRUS INCIDENCE IN THE COUNTRIES OF THE EUROPEAN REGION

Abstract

The article is devoted to topical issues of analysis and forecasting of the incidence of COVID-19 in the countries of the European region. From 2019 until now, the world has faced a new type of threat, namely the COVID-19 pandemic. The pandemic caused by this virus has already caused thousands of lives worldwide. Therefore, a detailed analysis of the level of morbidity will predict the further development of the disease, which in turn makes it possible to create more effective methods of combating the disease. The purpose of this article is to analyze and forecast the incidence of COVID-19 in the European region on the basis of available statistics obtained from the Internet resource Our World in Data, which is dedicated to solving global problems. To analysis and forecasting the level of coronavirus incidence, three countries with different levels of development were selected, namely: Germany, Spain and Ukraine. The authors reviewed methods, models and software tools for analysis and prediction of time series of incidence on COVID-19. During the choice of the model, the adaptive linear models of Holt, Box-Jenkins, adaptive seasonal models, autoregressive-moving average with a seasonal component were studied. It has been found that each approach and each method has its advantages, disadvantages and limitations in application. As a result, to forecasting the incidence of COVID-19, the ARIMA model, the autoregressive class of the integrated moving average, was chosen, as this model is quite flexible and allows taking into account seasonal fluctuations. To determine the optimal ARIMA coefficients of the model, Views software products and a set of Microsoft Excel add-ins for econometric analysis of time series were selected. The obtained results of prognosis indicate a stable course of coronavirus disease. But the obtained forecast values, in our opinion, are somewhat inflated, and the forecast error increased significantly with the forecast horizon. Therefore, this method of forecasting, based on the ARIMA model, can be further recommended for the construction of short-term forecasts of the incidence of COVID-19.