Travel Behavior Adjustment Based Epidemic Spreading Model and Prediction for COVID-19

Abstract

Due to the continuous variant of the COVID-19 virus, the present epidemic may persist in a long time, and each breakout displays strongly region/time-dependent characteristics. Prediction for each specific bursting is the basic task for the corresponding strategies. However, the refinement of the prevention and control measures implies generally the limitation of available records for the evolution of the spreading, which leads to a special difficulty for predictions. Taking into account of the interdependence of people's travel behaviors and the epidemic spreading, we proposed an Modified Logistic Model to mimic the COVID-19 epidemic spreading, to predict with limited epidemic related records the evolutionary behaviors for a specific bursting in a megacity. It reproduces successively the COVID-19 infected records in Shanghai China in the duration from March 1 to June 28,2022. Since December 7, 2022 when a new refinement of the prevention and control measures is adopted in the Mainland China, the COVID-19 epidemic blew up on a national-wide scale, and the drug "ibuprofen" is widely taken by the infected people themselves to relieve the fever symptoms. A reasonable assumption is that the total searching times for the word "ibuprofen" is a good representation for the amount of the infected people. By using the searching times for the word "ibuprofen" provided on Baidu, a famous searching platform in Mainland China, we estimated the parameters in the Modified Logistic Model and predicted subsequently the epidemic spreading behavior in Shanghai China starting from December 1, 2022. It will persist for a period of 72 days. The amount of the infected people will increase exponentially in the duration from the beginning to the 24th day, reach summit at the 31th day, and decrease exponentially in the duration from the 38th day to the end. Within the two weeks centered at the summit the increasing and decreasing speeds are both significantly small, but the increased amount of infected people each day is significantly large. The characteristics for this prediction match very well with that for the amount of metro passengers in Shanghai. As a proposal, the related departments should setup a monitoring system according to the principles of sampling in statistics, composing of some communities, hospitals, etc., to provide researchers with reliable records for prediction.