Using an Artificial Neural Network Model to Predict the Number of COVID-19 Cases in Iran

Abstract

Aim and background: Forecasting methods are used in various fields; one of the most important fields is the field of health systems. This study aimed to use the Artificial Neural Network (ANN) method in forecasting Corona patients in Iran. Method: The present study is descriptive and analytical of a comparative type that uses past information to predict the future, the time series of Corona infected patients in Iran was between May 2020 and May 2021. The forecasting accuracy index was the mean absolute percentage error (MAPE), and the forecasting method based on ANN was used to forecast the values ​​for June 2021. Input, output, and middle layer were used in designing the Neural Network (NN), and the training of the NN was based on the Levenberg-Marquardt algorithm. Findings: The mean absolute value of the designed NN error in forecasting the trend of Corona disease is 6%. ANN can forecast the course of the disease with 94% accuracy. Conclusion: ANN has good accuracy in forecasting the process of Corona disease. Furthermore, the outputs of this model can be used as a basis for decisions.