Assessment of dependence between first and second vaccinations with active COVID-19 cases in Jakarta, Indonesia through C- and D-Vine copula based regression
Falah Novayanda Adlin(a), Atina Ahdika(a*)

(a) Department of Statistics, Universitas Islam Indonesia
*atina.a[at]uii.ac.id

Abstract

Until now, Indonesia is still battling the COVID-19 virus, which has an increasing number of cases with an increasing varied type of virus. Jakarta is the province with the highest number of vaccinations 1 and 2. However, the number of active COVID-19 cases is still the highest in Indonesia. Based on these facts, there may be another dependency structure between the number of vaccinations and active cases of COVID-19 that ordinary linear correlations cannot capture. This study identified the dependence between the number of vaccinations 1 and 2 with the number of active cases of COVID-19 in Jakarta Province using the C- and D-vine copula models. The data used is time series data for the three variables from January 13, 2021, to September 24, 2021. Furthermore, an estimation of the number of active COVID-19 cases is carried out based on the number of vaccinations 1 and 2 with a vine copula-based regression model. The analysis results show that the dependence between vaccinations 1 and 2 is positive, between active cases of COVID-19 and vaccination 1 is positive, and between active cases of COVID-19 and vaccination 2 is negative. Finally, the best estimation results for active COVID-19 cases were obtained using a D-Vine copula-based regression model with a MAPE value of 0.597%.

Keywords: COVID-19, C-Vine copula, dependency, D-Vine copula, vaccination

Topic: MATHEMATICS AND STATISTICS