基于时变参数的SCUIR传播模型的构建与研究

Construction and Research of SCUIR Propagation Model Based on Time-varying Parameters

鉴于新型冠状病毒的传播特性,染病者不仅具有潜伏期且存在大量的无症状感染者,在经典SEIR模型的基础上,重新定义潜伏状态为密切接触状态,引入无症状感染状态,并考虑模型中状态转移参数会随着时间增长发生变化,提出了一种新的包含“易感状态,密切接触状态,无症状感染状态,确诊状态,移除状态”等五类状态的传播模型。利用湖北省真实疫情数据进行模型实验并对比结果,采用RMSE、MAPE值作为评价指标,结果表明,SCUIR模型的拟合精度有显著提升,与传统模型相比降低了8.3%~47.6%的拟合误差,并且可以计算出疫情中难以统计的隐藏数据,进一步刻画了疫情传播机理。

Novel coronavirus is a new type of virus,and its transmission characteristics are different from previous virus.Infected people not only have an incubation period,but also a large number of asymptomatic infections.Based on the classic model SEIR,this study redefines the latent state as close contact state,introduces an asymptomatic state of infection,and the influence of time on the state transition parameters in the model is considered,proposed a new transmission model which includes five types of states:susceptible state,close contact state,asymptomatic infection state,infected state,and removed state.The model uses the actual epidemic data of Hubei Province to conduct experiments,and uses RMSE and MAPE as evaluation indicators to compare the experimental results.The results show that the fitting accuracy of the SCUIR model has been significantly improved.Compared with the traditional model,the fitting error is reduced by 8.3%~47.6%,and hidden data that is difficult to count in the epidemic can be calculated,which further characterizes the mechanism of epidemic transmission.