Novel Coronavirus Forecasting Model using Nonlinear Autoregressive Artificial Neural Network
Novel Coronavirus (COVID-19) is considered one of the most significant public health theaters on the global level around the world. Due to the sudden nature of the outbreak and the infectious power of the virus, it causes people anxiety, depression, and other stress reactions. The prevention and control of the novel coronavirus pneumonia have moved into a vital stage. It is essential to early predict and forecast of virus outbreak during this disruptive time to control of its morbidity and mortality. This paper uses an artificial intelligence and deep learning for prediction of coronavirus through time series using Non-Linear Regressive Network (NAR). It can be considered as an effective approach to early warning and to aid politicians. The network is forecasted the coronavirus cases and deaths through nine countries namely; Egypt, Saudi Arabia, Jordan, United States of America, Spain, Italy, France, Iran, Russian Federation during the period from 23-Mar-2020 to 30-July-2020. All data is collected from the World Health Organization (WHO) situation reports and the confirmed cases and deaths are considered. It is found that training error in the network prediction which represents its regression and performance is 2.65 % in cases forecast and 3.22 % in deaths forecast for global prediction.