Public Opinion Evolution Law and Sentiment Analysis of Campus Online Public Opinion Events

• Published in: Journal of advanced computational intelligence and intelligent informatics 2024-07, Vol.28 (4), p.990-1004
• Main Authors: Xu, Zhengzhi, Ye, Zi, Ye, Haiyang, Zhu, Lijia, Lu, Ke, Quan, Hong, Wang, Jun, Gu, Shanchuan, Zhang, Shangfeng, Zhang, Guodao
• Format: Article
• Language: English
• Subjects: College campuses; Colleges & universities; Data analysis; Data mining; Digital media; Emotional factors; Evolution; Public opinion; Regional analysis; Regional development; Sentiment analysis; Social networks; Students; Teachers
• ISSN: 1343-0130; 1883-8014
• DOI: 10.20965/jaciii.2024.p0990

In the context of the new era, teachers and students in colleges and universities, as well as the general public, rely more on the Internet and social media to obtain news, express their opinions, and share information, and the dissemination of public opinion events in colleges and universities is not only related to the physical and mental health of teachers and students but also to the reform and development of colleges and universities. In this study, we took campus public opinion events as the main research object in which we selected three recent campus public opinion events to be analyzed. The public opinion data used in the research was collected from Weibo social media platforms. Firstly, we analyzed the dissemination cycle and regional dissemination patterns of a college food safety public opinion hot event through the popularity and regional distribution of public opinion data, thus revealing its formation and evolution patterns. Secondly, the LDA topic mining method is used to mine the themes of the three hot public opinion events, and then analyze the hot factors of the dissemination of each public opinion event from the massive public opinion data. This is crucial for the management department to grasp the dynamics of public opinion. Then, we used the SKEP sentiment classification method to analyze the emotional factors of the public opinion data of the three events to obtain the overall public opinion sentiment situation of the events. Finally, based on the characteristics of time, region, and gender, the evolution and diffusion rules of public topics and emotional distribution under different types of events are analyzed. The precision of the analyses associated with this paper may be limited to the effects of current mainstream as well as state-of-the-art analytical models. The analysis methods and conclusions in this paper provide a scientific theoretical basis and improvement measures for campus public opinion management, which helps to enhance the level of campus public opinion management and safeguard campus stability and public order.