Short Interval Forecasting Model of COVID-19 in IRAQ Using State Space Modelling

Dawood, A. A.; Al-Isawi, B. A.

In December 2019 COVID-19, a viral disease originated in Wuhan (China), spread through China, and break out through the rest of the world, the disease is called coronavirus. The world health organisation (WHO) declared this disease as pandemic in March 2020. The recommended measures in almost every infected country was to enforce social distance. In 17 March 2020 Iraq declared curfew “Social distance” and provinces lockdown for the entire country as a precautious measure to control the spread of the disease. The daily reports data collected from JHU can be represented as time series, thus it can be analysed using time series analysis methods. The challenges in this pandemic include the short in data points for each vital pointer, the current continuous exponential trend in confirmed cases, and the randomness as the dominant phenomenon in observations. Those reasons made it very hard to predict or forecast long periods in the future. This paper suggests short periods for forecasting and an online updated state space model. The model records high accuracy with low error percentages comparing with other statistical models, which indicate that this rank of predictors is more suitable than other statistical models. The proposed model based on state space model, which is a technique to forecast time series with unobserved parameter, seasonal and unseasonal. The exponential behaviour of the time series assists in the good performance of the predictors. © 2021, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.