Detection and analysis of emergency topic in social media considering changing roles of stakeholders

Abstract

PurposeMost of the existing studies on the evolution of emergency topics in social media focused on the emergency information demand of fixed user type in emergency while ignoring the changing roles of stakeholders during the emergency. Thus in this study, a three-dimensional dynamic topic evolution model is proposed, in which fine grained division of time, dynamic identification of stakeholders in the emergency, and emergency topic evolution based on both timeline and stakeholder's type are all considered.Design/methodology/approachParticularly the relevance between the tweets posted and the topic of emergency, the influence on the social network, and the attention of emergency topic are as well taken into account to quantitatively calculate the weight and ranking of stakeholders at different stages of the emergency. To verify the proposed model, an experimental demonstration was carried out under an emergency event posted on social media.FindingsThe results show that (1) based on the three-dimensional dynamic topic evolution model, the composition and ranking of stakeholders have obvious differences at different stages; (2) the emergency information needs and the sharing behavior of stakeholders on emergency information also indicate different preferences where the topic concerns of stakeholders at different stages have a strong relationship with their weight ranking; (3) the emergency topic evolution considering both the dynamics of emergency stakeholders and emergency information demand could more accurately reflect the changing regularity of social media users' attention to information in emergency events.Originality/valueThis study is one of first to investigate the emergency topic evaluation on social media by considering the dynamic changes of various stakeholders in emergency. It could not only theoretically provide more accurate method to understand how users share and search emergency information in social media, but also practically signify an information recommendation way in social media for emergency tracking.Peer reviewThe peer review history for this article is available at: https://publons.com/publon/10.1108/OIR-02-2021-0098.

Design of New Media Event Warning Method Based on K-means and Seasonal Optimization Algorithm

“…Fig. 11 Clustering results (6) The clustering results are composed into a classification model dataset, which is fed into the training RF classifier to obtain the new media event classification model. From the analysis of the above results, it can be seen that the Haimen incident belongs to the third category of new media events, which is a group event triggered by the normal demands of the masses, and there is a tendency to evolve to the fourth category of new media events.…”

“…Timely and effective early warning of new media events not only provides academic value to the study of new media events, but also can play a positive role in promoting the resolution of public opinion [5]. Therefore, the research on new media events based on network technology is extremely necessary [6].…”