Analysis And Mining Of The Evolution Process Of Network Public Opinion And Application In Teleco-Mmunications Fraud Events

In recent years,along with the popularity of network equipment,the explosive growth of network information,the public with the help of the network platform to express their views on the event,and finally form the network public opinion on the event,network public opinion also has an impact on the public’s views,attitudes and emotions.Through the study of the evolution process of network public opinion,we can find the emotional changes and focus of the people’s concern sitts for hot events,and provide reference for the supervision and management of the network public opinion of the relevant departments.In this paper,we choose the typical event "Xu Yuyu Telecom Fraud Case" which has aroused the enthusiastic response of the whole society in recent years to study the emotional changes and focus in the evolution of public opinion in the network.A research method on the evolution of network public opinion combining text subject analysis,text emotion analysis and semantic analysis is proposed,and the evolution stage of network public opinion is analyzed from many angles.By analyzing the text data of news reports and comments related to hot events,the evolution of public opinion of hot events is divided into three stages: outbreak period,transition period and smooth period according to the growth of the number of news reports.By comparing the changing trends of news articles with reviews,we find that the news media and the public’s attention are growing in the period of outbreak and stability.During the turning period,the news media is consistent with the trend of public concern growth,and the latter growth trend is the opposite,influenced by the information about emerging events.In view of the three stages of the evolution of public opinion,the Sentence-LDA model is introduced to analyze the subject of news reporting.We select the emotional analysis model based on the emotional dictionary and the semantic network model to analyze the comment data of the three stages of news reporting based on the emotional dictionary,and explore the emotional tendency and focus of the public’s exposure to hot events.The results of the three period analysis of the evolution of public opinion by Sentence-LDA theme model show that the focus of the report shifts from focusing on the description of the case to the specific details of the case,and finally extends to the cause of the incident and the measures to solve it.The results of the emotional analysis model show that the tone of public comments on negative events is negative,with 92.52 percent negative emotions and 7.22 percent positive emotions.Emotional polarity and intensity gradually fall from the peak,and eventually fluctuate in a small range.The semantic network model results show that the public’s focus is basically the same,but the specific content of the discussion around the focus will change with the evolution of public opinion.This combination of different stages of evolution of network public opinion and a variety of text mining research methods can reduce the loss of data in the process of text mining,a comprehensive observation of the network public opinion evolution of different stages,for the relevant departments to the network public opinion development trend of the judgment to provide a theoretical basis.