Forecasting of COVID-19 Using Modified SEIR, Logistic Growth and Holt’s Models

Mahmoud, T. M. A.; Mohsen ElOcla, N.; Abu-Tafesh, M. S. A.; Mohamed, A. S. A.

Understanding the transmission dynamics of the novel coronavirus is the concern that attracted many researchers nowadays. In this chapter, three mathematical models, modified SEIR, logistic growth and double exponential smoothing, were implemented in Matlab to predict the transmission of COVID-19 in Egypt, Oman, Nigeria and South Africa. To estimate the models’ parameters, the reported data were used to fit the models using Nelder-Mead, Levenberg-Marquardt, Trust-Region-Reflective, and MultiStart object with Interior-Point optimization algorithms. Then, a sensitivity analysis was made to understand the effect of different parameters on the models’ prediction. The application of the three models on the reported data was compared despite their different nature. It was shown and verified that the three models are highly dependent on the parameters’ values, referring to the importance of determining their estimates using an optimization algorithm. It was found out that the most dominant parameter is the one denoting the rate by which susceptible people are protected, which emphasizes the effect of social distancing and quarantine. © 2021, The Author(s), under exclusive license to Springer Nature Switzerland AG.