Fine-grained detection on the public's multi-dimensional communication preferences in emergency events.

With the rapid development of Internet technologies, the public can participate in the information communication of emergency events more conveniently and quickly. Once an emergency occurs, the public will immediately express and disseminate massive information about the causes, processes and results of the emergency. In the process of information communication, the public often adopts diversified communication modes, and then shows differential communication preferences. The detection of the public's communication preferences can more accurately understand the information demands of the public in events, and then contribute to the rational allocation of resources and improve the processing efficiency. Therefore, this paper conducted finer-grained mining on the public's online expressions in multiple events, so as to detect the public's communication preferences. Specifically, we collected the public's expressions related to emergency events from the social media and then we analyzed the expressions from multiple dimensions to obtain the corresponding communication features. Finally, based on the comparative analysis of diversified communication features, static and dynamic communication preferences were obtained. The experimental results indicate that the public's communication preferences do exist, which is universal and consistent. Meanwhile, constructing a better social environment and improving people's livelihood are the fundamental strategies to guide public opinion.