Reasearch On The Opinion Evolution And Sentiment Analysis Of Public Opinion Reversion

With the rapid development of the Internet,the self-media represented by We Chat and Weibo have become the main media for information dissemination.Public opinion has shifted from the real world to the virtual space of the network.Some social emergencies or public issues often become the focus of public opinion and induce network public opinion.Public opinion reversion is an important phenomenon of network public opinion.The frequent occurrence of reverse events not only damages the credibility of the official media,but also causes negative social effects.At present,researches on public opinion reversion mainly focus on qualitative analysis,only few studies analyze the evolution of public opinion reversal from a quantitative perspective.Therefore,it is significant to use scientific and reliable technique in analyzing the causes and evolution of public opinion reversion.Based on the individual of real society,this article combine mathematical modeling,computer simulation analysis,and case data verification to study the evolution of public opinion reversion.Firstly,in order to analyze the process of opinion evolution of public opinion reversion,we propose a model based on opinion leaders from the perspective of opinion dynamics.Considering factors such as individual heterogeneity,influence of opinion leaders,and socially complex networks,the model of public opinion reversion is constructed on the basis of the classic bounded confidence HK model.We use the central indicators of social network analysis to identify opinion leaders,conduct simulation experiments on opinion evolution under the guidance of opinion leaders,analyze and discuss the experimental results.Secondly,considering the characteristics of short text,we propose a sentiment analysis method based on deep learning,use the word2 vec word vector to represent the text,integrate Bi GRU,attention mechanism and capsule network to build a hybrid network model.We also use the Chinese short text dataset to train and test the model,then design comparative experiments from multiple dimensions to verify the effectiveness of the model.Finally,we select a public opinion reversal case on Sina Weibo for the research,the case data was sorted and analyzed using opinion dynamics modeling and text sentiment analysis technology,by comparing the experimental results and the case data we analyze the evolution process of public opinion reversion.This paper focuses on modeling the interaction rules of individuals,mining emotion information of the text,and conducting research on the evolution of public opinion reversion.The experimental results and case study verify the effectiveness of modeling and sentiment analysis methods.The research results provide a reference for relevant departments to formulate opinion guiding strategies,which have important academic value and practical significance.