Generalized logistic growth modeling of the COVID-19 outbreak in 29 provinces in China and in the rest of the world

Ke Wu; Didier Darcet; Qian Wang; Didier Sornette

Since it is delicate to interpret the absolute number of infected cases, we propose to rather focus on the dynamics, which is more informative as we have presented above. For South Korea, we have shown that it is approaching the ceiling in the total number of infected cases and we do not expect this number to grow significantly in the coming future. If we assume a generalized logistic model for the behavior of Italy, the estimated final fraction of population infected will be 0.15% (95% CI: [0.03%, 0.30%]), which will be much higher than Hubei. reporting more than 1000 confirmed cases. The statistical analysis shows that the epidemic in Europe is in the early exponential regime and this may continue for a while. At the time of writing, the estimates for the tapering away from a pure exponential growth and the transition to an inflection point and a decay, as obtained from both generalized logistic model and logistic model, are not reliable. There are huge uncertainties simply stemming from the fact that the growth is very close to being pure exponential and does not reveal much additional information.