Modeling the Number of COVID-19 Total Cases in Iran Using Gompertz and Logistic Growth Curves
Abstract
Introduction: The growth curve are time dependece regression models which commonly are useful in describing the rapid growth of total cases or deaths in a pandemic situation.
Methods: The Gompertz and logistic functions are useful to describe the growth curve of a population or any time dependence variable such as metabolic rate, growth of tumors and total number of cases or deaths in a pervasive disease. The logistics family of growth curve including logistic, SSlogistic, generalized logistic and power logistic and Gompertz models were considered to describe the growth curve of total_cases_per_million (t_c_p_m) of COVID-19 in Iran during the 19-Feb-2020 to 28-May-2021. The models were fitted to data using nls function in R and the fitting accuracy was evaluated using the numerical and graphical approaches.
Results: The logistic family and Gompertz growth curve were applied to fit the total_cases_per_million of COVID-19 in Iran as the response versus the time in days as predictor variable. The training and testing RMSE criterions were considered as the numerical criterions to assess the model accuracy. The growth curve of fitted models was compared with the growth curve of observed data. Results indicated that the logistic and Gompertz models provided a better description of target variable than the alternatives.
Conclusion: As results shown, the logistic and Gompertz models provided a better description of response variable than the alternatives. Therefore, the logistic and Gompertz models are able to describe and forecast the COVID-19 variables (including total cases, death, recovered and so on) very well.