Using Some of the Nonparametric Smoothing Models to Build An Appropriate Model for Predicting Number of Cases of Covid-19 in Iraq

Abstract

Using nonparametric smoothing models is more effective for the data than the models that depend on the prior functional formula of the data, and the most important of these models are the Exponential Smoothing models. The most important models used in exponential smoothing will be discussed, including Holt-Winter's Models (HWM) and based on some criteria, the best suitable model will be chosen for forecasting. Also, Box & Jenkins model will be used in the analysis of the time series to find the best prediction model. The importance of research comes from the problem of the spread of the epidemic (covid-19) all over the world, as well as in our country, Iraq, which led to almost a complete halt to all life facilities. The aim of the research was to use the methods mentioned in the case of forecasting time series and to choose the most appropriate model for predicting the number of infections of the COVID-19 epidemic in Iraq. The sample included the number of infections of the daily epidemic in Iraq for the period (1/6/2020 to 1/6/2021). The results showed that the Multiplicative Seasonality Model is appropriate according to the criteria of the Holt-Winter's Model (HWM), and the appropriate model according to the method of using the Box & Jenkins model is ARIMA(2,1,3). These models predicted the number of upcoming infections in a subsequent month period (30 days).