Constructing an infectious disease nowcasting model based on data assimilation technology:taking COVID-19 as an example
Based on the traditional dynamical SEIR model,control and isolation measures as well as the protective role of vaccines were considered to reconstruct the background SE1QRDV model.Furthermore,the assimilation model for epidemic nowcasting,SEIQRDV-EnKF,was established by integrating the Ensemble Kalman Filter(En-KF)technique.The model's performance was evaluated using COVID-19 data from Hubei Province,China,the U-nited States,and India.The results showed that the predicted number of infections and recoveries by the assimila-tion model closely matched the actual data,with low prediction errors in terms of root mean square error(RMSE)and mean absolute percentage error(MAPE).Moreover,it overcame the limitations of the dynamical model and ac-curately predicted the epidemic trend using relatively short historical data.This provides strong technical support for local government decision-making and deployment during major infectious disease outbreaks.
万方数据
