兰州理工大学机构知识库

Twitter can play an important role in the control of influenza epidemics. We introduce a quantitative approach to evaluate the effects of Twitter on the modeling of the spread of influenza epidemics in this paper. Statistically significant correlations between the number of the percentage of tweets that are self-reporting flu and data of influenza-like illness reported cases are found from Pearson correlation and cross-correlation analyses, during the 2009 H1N1 flu outbreak in England. A new H1N1 model with relapse which involves impact of Twitter are also proposed. Stability of all the equilibria of our model are obtained. The occurrence of backward and forward bifurcation are also established. The best-fit parameter values in our model are identified by gray wolf optimizer and nonlinear least square method from the above data. For determining key parameters during the outbreak of the disease with Twitter impact, the uncertainty and sensitivity analyses are explored by using a Latin hypercube sampling (LHS) method and evaluating the partial rank correlation coefficients (PRCCs). Our results show that Twitter reports have important implications for the control of infectious diseases and Twitter can serve as a good indicator of influenza epidemics. © 2020 Elsevier B.V.