Forecasting the Number of (Infected, Recovered, Death) For COVID-19 in Iraq by Using Neural Network Models

Abstract

Predicting the diseases caused by viral infections is a complex medical task as many real data consisting of different variables must be used. As it is known, COVID-19 has created a great crisis in all countries of the world because of the repercussions it has left and will leave behind, it is necessary to find a way to explain the spread of this virus by relying on data of many infected, deceased and recovered people, so artificial neural networks were used ( ANN) to predict the numbers (infected, deceased, recovered) of the COVID-19 pandemic, because the idea of ??\u200b\u200bthe work of neural networks is the process of simulating data to reach a model for this data for the purpose of analysis, classification, prediction or any other treatment without resorting to a proposed model for these data, as An appropriate neural network was designed, as well as an automatic algorithm  for teaches and trains the network to reduce the error coefficient to the lowest level, Forecasting was also carried out in nonlinear models, which proved that neural networks are fully capable of predicting. One of the most important conclusions reached by the research is the appropriateness of the Radial Basis Function Neural network (RBFNN) model to modeled and to forcasting of the data of the numbers of infected, deceased and recovered patients for the Corona virus in Iraq for the period from (1/22/2020-31/7/2021) and this is what was clarified by the values of MSE, RMSE AND R2, addition that The predictive values were very close to the true values, and this confirms that the RBFNN model is very suitable.