Time series forecasting of COVID-۱۹ in Iran using Holt-Winters smoothing methods

The rising trend of COVID-۱۹ has put the world economy and the healthcare system of countries over the world in the crisis. At present, the shortage of healthcare facilities has been emerging as an urgent challenge. Designing strategic and operational decisions for hedging against the risk of this epidemic requires an estimate of the new cases of the infection in the future. This paper presents different forecasting methods to predict the time series of COVID-۱۹ in Iran. To forecast the daily and cumulative fluctuations of new cases, several Holt-Winters exponential smoothing methods are used. These methods include the additive Holt-Winters method, the extended Holt-Winters method, and the additive Holt-Winters method with damped trend. The methods are tested on the data series of Iran. Different efficiency measures are also presented to evaluate the performance of forecasting methods. The empirical results show that the extended Holt-Winters method leads to more accurate forecasts. Moreover, the forecasted values obtained with other methods are reliable and satisfactory, especially in forecasting the cumulative cases of COVID-۱۹.

کلیدواژه ها:

COVID-۱۹ ، Forecasting ، Time series ، Exponential smoothing ، Holt-Winters methods ، Performance measures