Research On The Theme And Emotion Evolution Of Internet Public Opinion On Hot Events Oriented To Multi-agent Collaborative Governance

The rapid popularization of the Internet has caused people to spend a lot of time surfing the Internet.Therefore,whether the network environment that people often come into contact with is healthy and orderly has become a core issue of people's attention.Among them,network public opinion is an important aspect of whether the network environment is healthy.At present,the research direction of online public opinion is diversified,and the research in various vertical fields is diversified but there is little cooperation.There is no multi-angle collaborative analysis of problem phenomena.Although online public opinion is spread and fermented online,it has a huge impact on all aspects online and offline,and it is related to social stability and security.In the development of online public opinion,the public's emotions are divided into six categories: good,sad,fearful,frightened,angry,and evil.Their emotional changes are extremely important and sensitive,affecting the overall emotional development of the event,and thus have a certain impact on social stability.Therefore,it is very important to mine and analyze the emotions in the evolution of online public opinion.The specific work is as follows:This research proposes a framework for the evolution of popular event topics and emotions for multi-agent collaborative management.The framework is divided into three modules: data processing,data modeling,and data analysis.Determine the research data platform and source,and select Micro-blog as the data platform for this research based on hot events.First,the data processing module includes the use of a data searcher to crawl experimental data,including Micro-blog texts and corresponding comments and related blog information,and the use of python to preprocess the data to form a Micro-blog corpus.On this basis,divide the life cycle stages of multiple governance entities and hot events.Secondly,the data model of this study is constructed and divided into a text topic mining model based on neural network and an emotional computing model based on sentiment dictionary.The Micro-blog text corpus uses the topic clustering methods of Word2 vec and Kmeans to classify topics according to the different life cycle stages of different partners,and then uses the sentiment dictionary method to calculate sentiment categories and different topics,different stages,and reviews under different topics.This score is used for sentiment analysis to explore the impact of themes proposed by different themes on public sentiment,so as to obtain the best governance method for each governance theme at different stages of evolution.Finally,taking the "Hangzhou Missing Women Case" as an example,an empirical analysis of the theme clustering and emotional evolution of each governance subject in the life cycle of a specific social hot event is carried out.In hot events,the theme mining is carried out on the different life cycle stages of each governance body,the reasons for the theme changes and the public sentiment changes caused by the theme changes are studied,and the governance timing and governance effectiveness have been verified by each governance organization.Ways to participate in the evolution of public opinion.This research focuses on the key issue in cyberspace governance-the governance of Internet public opinion on social hot events,and deepens the theoretical system of cyberspace governance.Through thematic analysis and sentiment analysis methods,the monitoring of the evolution of social hot public opinion is realized,and the collaborative governance model is integrated,which provides new research perspectives and research ideas for Internet public opinion research.At the same time,it also helps meet the current urgent needs of online public opinion governance,and provides practical public opinion monitoring and collaborative governance methods.