Adaptive Propagation Model of Network Hotspot Events Based on SEIR

Abstract

Internet hotspot events spread quickly and have a significant influence on the Internet, becoming the focus of monitoring public opinion. Due to the gradual fermentation of these events, the scope of transmission, the number of participants, and the event's influence constantly change. Therefore, a propagation model with fixed parameters cannot accurately describe the propagation law of hotspot events. To address these issues, this paper proposes an adaptive SEIR propagation model, called SEIR-A, which incorporates a dynamic infection rate. This model enhances the traditional SEIR model by considering susceptible, latent, infected, and cured individuals. Furthermore, it combines the model with the data assimilation method to capture the trend of hot topics. Experimental testing demonstrates that the model effectively describes the propagation trend of four network hotspot events from Weibo, thus proving its accuracy and applicability.