Application of genetic algorithm combined with improved SEIR model in predicting the epidemic trend of COVID-19, China

Abstract

Since the outbreak of the 2019 Coronavirus disease (COVID-19) at the end of 2019, it has caused great adverse effects on the whole world, and it has been hindering the global economy. It is ergent to establish an infectious disease model for the current COVID-19 epidemic to predict the trend of the epidemic. Based on the SEIR model, the improved SEIR models were established with considering the incubation period, the isolated population, and genetic algorithm (GA) parameter optimization method. The improved SEIR models can predict the trend of the epidemic situation better and obtain the more accurate epidemic-related parameters. Comparing some key parameters, it is capable to evaluate the impact of different epidemic prevention measures and the implementation of different epidemic prevention levels on the COVID-19, which has significant guidance for further epidemic prevention measures.

Prediction of cross-border spread of the COVID-19 pandemic: A predictive model for imported cases outside China

“…Since the first outbreak of COVID-19, predicting its development trend has never been stopped, and a number of prediction models have been proposed, such as the combined prediction models [ 26 ] based on complete ensemble empirical modal decomposition (CEEMDAN), extreme gradient elevation tree (XGBoost), and network search data (WSD). Also, methods [ 27 ] for predicting the development trend of COVID-19 have been attempted, such as the long- and short-term memory (LSTM) neural network of Dropout technology and the SEIR optimization model [ 28 ]. However, CEEMDAN, XGBoost and WSD are complex and require high timeliness of the data [ 26 ].…”

“…For the LSTM neural network approach to predict the development trend of COVID-19 [ 27 ], the error interval is too large when it is used to predict the cumulative confirmed numbers based on a large base number within a long statistical time. Although the SEIR model considers the impact of novel corona virus infection latency on the development of the epidemic, the latency individuals in the actual transmission process, the difference in the onset time is difficult to count [ 29 ], and objective factors such as isolation measures have a greater impact on it, resulting in that the prediction model cannot accurately obtain the real epidemic parameters [ 28 ]. In the actual spread of the epidemic, it is reasonable to assume that the infection development and change of the infected persons and latent persons in a country can be finally reflected in the total number of the confirmed persons in that country, based on which a better development-predicting outcome can be obtained, this assumption gets support from Kremer [ 30 ].…”

Difference in Rumor Dissemination and Debunking Before and After the Relaxation of COVID-19 Prevention and Control Measures in China: Infodemiology Study

“…For instance, Xi and Yong [18] proposed a model for public opinion dissemination on social networks, Han et al [19] analyzed network topology in rumor spreading, Wu et al [20] developed a rumor-spreading model based on recommendation systems, and Fang and Yang [21] studied rumor spreading and refutation effectiveness on Weibo under real-name systems. Moreover, epidemic models have been applied to social network rumor propagation studies, based on similarities between rumor spread and disease transmission mechanisms [22], including SI (susceptible and infectious) [23][24][25], SIS (susceptible, infectious, and susceptible) [26][27][28], SIR (susceptible, infectious, and recovered) [29][30][31], and SEIR (susceptible, exposed, infectious, recovered) [32][33][34] models, representing susceptible, infected, recovered, and exposed states, respectively [35]. Zhu et al [36] introduced a SIS model with a forced silence function, Sun et al [37] applied uncertain differential equations to study an SIR model, Hosseini and Zandvakili [38] added rumor delay and countermeasure mechanisms to an SIR model, and Chen et al [39] proposed a new SEIOR (ignorant, hesitators, spreaders, rumor debunkers, and immunizers) model incorporating debunkers.…”

STG-Net: A COVID-19 prediction network based on multivariate spatio-temporal information

“…In mathematical epidemiological models, most studies have analyzed the spread of COVID-19 by modifying and adapting two models, SIR (Susceptible, Infected, and Recovered) and SEIR (Susceptible, Exposed, Infected, and Recovered) [2] , [3] . Liao et al [4] proposed a time-window-based SIR prediction model that dynamically analyzes data through time windows and then uses machine learning methods to predict the underlying reproduction number and exponential growth rate of the epidemic.…”