Abstract
Short videos are gaining popularity among Chinese netizens who use them to socialize and express themselves online, particularly during disasters. This study examines the characteristics of an emotional atmosphere through short video comments. We developed the Multimodal Social Short Video Crawler V1.0 to analyze 157,747 comments from 343 short videos about the “Zhengzhou flood.” First, we categorized comments into clusters and calculated cluster density using complex network analysis methods. Second, we employed computer-mediated content analysis to derive sentiment values in clusters. To capture the emotional atmosphere in short videos accurately, we developed a trust model to assign weights to both cluster density and sentiment value. Our findings reveal that the emotional atmosphere in short video comments indicates continuity and clustering traits. Furthermore, a correlation was found between cluster size and emotional valence, with the relationship trend varying between macro- and micro-clusters. Regarding theoretical contributions, this study enriches the theory of emotional atmosphere by proposing a mathematical model of emotional atmosphere that combines machine learning and complex network analysis, and introduces cluster density and emotional value calculation methods. Concurrently, this study highlights emotion’s importance in short video comments, which can enhance disaster response efforts in the future.
Disclosure statement
No potential conflict of interest was reported by the authors.
Notes
1 1. To better visualize the distribution of emotion clusters in short videos, some comment nodes were extracted in the drawing medium ratio and displayed in the .
Additional information
Funding
Notes on contributors
Xiaohong Wang
Xiaohong Wang, professor, dean of the Undergraduate School, Communication University of China. Her primary research areas include broadcast journalism, online video and new media, and visual political rhetoric.
Chen Zhang
Chen Zhang, a Ph.D. student at Communication University of China majoring in Internet information, conducts research on information communication on social media. Her work has examined social media’s role in post-disaster recovery, as well as public opinion and emotions during disasters.
Qinglan Wei
Qinglan Wei (Member, IEEE) is currently a postdoc researcher in the School of Data Science and Intelligent Media, Communication University of China. She has published articles in academic journals and conferences. She has twice received the Second Runner-up position of the Group Emotion Recognition Sub-challenge sponsored by ICMI, ACM. Her main research interests include machine learning, computer vision, and emotion analysis.
Yichun Zhao
Yichun Zhao is a postgraduate student majoring in international journalism and communication in the School of Television at the Communication University of China. Her research focuses on social interactions in cyberspace.