Research on the Evolution of Public Opinion Themes on Microblogs of Emergency Events Based on the BERTopic: A Case Research of the Eastern Airlines Flight MU5735 Crash

Subjects: Library Science，Information Science >> Information Processing submitted time 2024-04-18

Jiang Changbin Xu Sisi Huang Yinghui Xu Ziqi Wang Hongyu

Abstract： Purpose/Significance : This research aims to systematically analyze the thematic evolution trends of public sentiment during emergency events, visualizing the focal themes throughout the development process of public sentiment, and providing a practical reference for guiding future online public sentiment. Method/Process : Utilizing the BERTopic model for topic extraction, this research identifies the themes at different stages of sentiment development and measures the similarity between themes using cosine similarity to visualize the thematic evolutionary paths. The case of the Eastern Airlines Flight Mu5735 crash on Sina Weibo is examined to research the thematic evolution of public sentiment during an emergency. Results/Conclusion : The empirical results demonstrate that the BERTopic model is effective and offers high visibility in theme identification for sentiment events, accurately capturing the hot topics in each phase of sentiment development and revealing the thematic evolution process during the spread of public sentiment. Innovation/Limitation : In this research, we propose a general framework for analyzing the theme evolution of microblog public opinion on short-text emergencies based on the BERTopic model, and we analyze the theme content evolution of the extracted results and present them visually. The limitation of this study lies in the fact that the data sources selected in this study are only from the microblogging platform, and the diversity of data sources can be improved in the future.

Peer Review Status:Awaiting Review

 Hits 194  Downloads 78  Comment