Research on Infectious Disease Prediction Method Based on Dynamic Mobility Network and Segmentation Algorithm

The high contagiousness of COVID-19 poses a serious challenge to the global public health system. This article aims to explore how to effectively predict and scientifically evaluate the spread of infectious diseases under the continuous adjustment of epidemic prevention measures using population mobility networks and segmentation algorithms. The article first integrates a dynamic mobility network that reflects the characteristics of population mobility into an epidemiological model. The nodes in the mobile network represent geographical regions within the city, and the edge weights represent population mobility between different regions. This model fully considers commonly used epidemic prevention measures. Subsequently, considering the transmission characteristics of infectious diseases and the variability of measures, we designed a new segmentation algorithm. Specifically, we first use the Chow test to automatically obtain data breakpoints, then adjust breakpoints based on changes in parameters and population mobility data within adjacent segments, and use least squares to solve for the coupled propagation parameters of each region, in order to improve prediction accuracy and identify policy equilibrium points. Finally, the superiority of the method is verified by experiments.