Statistical Exploration and Prediction of the Total Number of Daily Cases and Deaths of Coronavirus Disease 2019 (COVID-19)

Md. Mamun Miah; Jahid Hasan

Abstract

The study deals with novel coronavirus disease (COVID-19) completed here in two parts. First, we discussed some epidemiological parameters and their comparative study among some highly affected countries. In the second, though random phenomena (like cases and deaths) are unpredictable we tried to find time series ARIMA model and to forecast the nationwide daily total cases and deaths using worldometer database. From January 22 to March 18, 2020, a total of 219240 Coronavirus cases 9% of patients died and 91% of patients have recovered. Besides this, among a total of 219240 currently infected cases, 37% in China and 63% cases are outside of China of the world. To test the stationarity graphical method is used here. The data have been differenced twice to convert from non-stationary to stationary. From the autocorrelation function (ACF) and partial autocorrelation function (PACF) we obtain the order of the time series model. The chosen model was ARIMA (1,2,2) for the daily total number of cases and ARIMA (2,2,5) for the daily total number of deaths. These models have been fitted on data to estimate the parameters of autoregressive and moving average components of ARIMA (1,2,2) and ARIMA (2,2,5) models. For residual diagnostics, histogram and normality tests were used. Using model selection criterion and checking model adequacy, we see that the model is suitable in shape. It is found that the daily total cases and total deaths are increasing day by day. If no effective drugs or vaccine is made immediately more and more people will die of COVID-19.

Abstract

Keywords