Estimation and prediction of the prevalence rate of COVID-19 disease based on multilayer perceptron artificial neural networks model

Background & Aims: Nowadays, with the coronavirus disease-2019 (COVID-19) pandemic, millions of people have been infected with the coronavirus, and most countries in the world have been unable to treat and control this condition. The aim of this study was to estimate and predict the COVID-19 prevalence rate based on multilayer perceptron artificial neural network (MLP-ANN) model. Materials & Methods: In this cross-sectional study, based on the information of 4,372 patients with COVID-19 referred to Dr. Masih Daneshvari Hospital in Tehran, the prevalence rate of this disease was estimated. In addition, considering the role of the health measures and social restrictions, the trend of this index based on the MLP-ANN model was predicted. Results: According to the results of this study, the prevalence of COVID-19 increased by an average of 7.05 per thousand people daily during the 48 days from the onset of the epidemic, and it reached about 341.96 per thousand people. Based on the MLP-ANN model with a lack of attention to the health measures by individuals in the community and failure to reduce social restrictions by the government, the COVID-19 prevalence increased by an average of 1.03 per thousand people per day. While in the case of attention to the health measures by the people and continued social restrictions by the state, the prevalence of this disease decreased by an average of 2.13 per thousand people, daily. Conclusion: The study on the prevalence of COVID-19 disease and prediction of the trend of this index provides researchers with useful information about the role of the health measures and social restrictions in controlling this disease.