COVID-19 Highest Incidence Forecast in Russia Based on Regression Model

Abstract

The authors suggest a simple regression model of COVID-19 highest incidence prognosis in Russia on the basis of the revealed correlation between the duration of coronavirus peak (plateau) and air traffic volume. The study base included 37 countries in Europe, South America and Asia. Cluster analysis on the basis of the Euclidean metric for these countries showed the necessity of classifying the USA and China into a separate group, which gave grounds to exclude these countries from the analysis. In addition, Ireland was excluded from the analysis due to its special geographical location. For the remaining countries, the correlation coefficient between the number of airline passengers and the duration of the epidemic before reaching its peak was 0,87, which shows a high level of linear relationship between these indicators. Point forecast for the highest incidence in Russia by regression line falls on the 4th of May. The forecast interval with confidence levelγ=0.9 is ±14 days from the calculated date. The one-way analysis of variance showed that from April 22 to May 2, there was a slowdown in the growth rates of the diseased, which indicates an exit to the plateau.

Keywords

• covid-19
• regression model
• forecast
• peak (plateau)