A New Model for Epidemic Spreading with a Focus on COVID-19

Background: It is extremely useful to construct mathematical models to forecast and control real phenomena. One of the common applied statistical models to represent the data involving with time is the time series modeling. A novel time series model to represent the propagation of an epidemic infection in a population is presented. The model deals with addressing the cumulative number of confirmed cases. Methods: Our model is the generalization of statistical exponential growth models and can describe different stages of the outbreak of a communicable disease. Applying the mentioned procedure leads to models CVJR1 (3.2, 1.44, 3, 13) for modeling the sequence of COVID-19 from January 13 to March 5. All computations and 200 simulations were done in MatLab 8.6. Results: For comparing candidates through fitting the dataset for six pairs of ((l) over cap and (a) over cap), we used the minimum criterion square of residuals. We present the average and 90% upper and lower bounds of the predictions made by our models for three periods. Applying the mentioned procedure led to having models with parameters (3.2, 1.44, 3, 13) for modeling the course of COVID-19 from January 13 to March 5. Conclusions: The presented model can cover the epidemic behaviors related to social networks. Our model can be adjusted to worldwide modeling for modeling a phenomenon spreading in different populations simultaneously.