Analysis of COVID-19 Dynamics in EU Countries Using the Dynamic Time Warping Method and ARIMA Models

Abstract

AbstractThe aim of the paper is to find the similarities in the evolution of time series for people infected with and died from COVID-19 in different EU countries using dynamic time warping (DTW) as a measure of the distance between time series. Using this method, a joint analysis of the number of infected and deceased will be performed. The DTW distance makes it possible to compare time series of different lengths, which is important when analyzing data for European countries because the virus has not spread to individual countries at the same time. After measuring the similarities between the time series, a hierarchical grouping for countries will be performed, which will allow us to find interesting patterns in the data. Then, ARIMA(p,d,q) models will be used to describe the dynamics of virus distribution in different EU countries. With these models, it is possible to gain knowledge about the mechanisms of pandemic evolution.KeywordsCOVID-19Dynamic time warpingHierarchical clusteringARIMA models