| At present,China is in a critical period of deep social transformation and economic transition.With the rapid development of China’s economy,the process of new industrialization and economic globalization will be further accelerated.During the transformation period,not only social progress and development but also social contradictions become increasingly prominent,leading to frequent emergencies.In the context of the rapid development of the Internet,the Internet has become the main way for people to spread information and make comments.Once an emergency occurs,it will attract the attention and discussion of a large number of netizens in a very short time.They will express their personal emotional attitudes and opinions on the network platform,which makes the information of public opinion grow rapidly and spread rapidly.It is easy to cause public opinion waves in emergencies.If relevant departments do not guide and control it,it may cause social unrest.Compared with traditional public opinion,online public opinion spreads faster,affects more people and causes more serious consequences.As the main participants in discussion,the emotions and focus of Internet users can reflect the trend of public opinion to a certain extent.Therefore,by studying the evolution of Internet users’ emotions and the characteristics of the themes they pay attention to,It can enable the government to grasp the evolution of online public opinion in emergencies in time and conduct appropriate guidance and intervention at key nodes,so as to avoid the negative impact on society caused by the uncontrolled development of the situation.This paper proposes a research framework of online public opinion for emergencies,which integrates OC-Bert text sentiment analysis and LDA text topic analysis,so as to analyze the emotional evolution and theme characteristics of online public opinion for emergencies.On the one hand,OCC affective cognition and BERT deep learning algorithm are used to achieve emotion classification of online public opinion texts in emergencies.Experiments are designed to compare OCC-BERT with other emotion classification methods.The results show that BERT’s emotion classification model marked by OCC has the highest accuracy.It is proved that it can well complete the emotion classification of emergency online public opinion text.On the other hand,on the basis of emotion classification model,an LDA-based topic analysis and research framework of online public opinion for emergencies is proposed.In this topic analysis,the characteristics of stage division of public opinion,emotion classification results and topic mining results are integrated to analyze the topic characteristics of online public opinion for emergencies.This chapter mainly takes "Tianjin Explosion on August 12" as a specific case,and makes an empirical analysis of OC-Bert sentiment classification model and LDA topic mining proposed above.Based on data acquisition and processing and evolution cycle division,emotion evolution analysis,topic mining analysis and topic feature evolution analysis are carried out.Through empirical analysis,the hot topics and emotional evolution that netizens pay attention to in different stages of the development of public opinion are dug out.Mastering the development direction of online public opinion and the attitude of netizens in the 8.12 Tianjin explosion emergency is conducive to the relevant management departments to obtain the development trend of public opinion and the emotional changes of the public.And provide relevant departments with effective evolutionary analysis path for online public opinion in emergencies and policy suggestions for public opinion guidance. |
PDF Full Text Request