Research On Early Warning Of Public Opinion On The Background Of Social Stability

The rapid development of China's Internet at this stage makes Internet public opinion gradually become an important factor affecting social stability,especially the increase in the number of crises caused by Internet public opinion in recent years,making the need for early warning of Internet public opinion imminent.In the context of social stability,this article uses a variety of research methods such as literature research,qualitative and quantitative research methods,and case studies to build an Internet public opinion early warning index system,and clarifies the important index factors that affect the level of online public opinion early warning in the context of social stability.At the same time,a comprehensive and reasonable online public opinion early warning model was constructed to provide general ideas and specific suggestions for the establishment of China's online public opinion early warning mechanism,thereby assisting the public opinion supervision department to discover social hotspot information or negative information in a timely manner and make reasonable early warnings for maintaining social stability.This article starts with the study of the main concepts and basic theories.First,it defines the basic concepts of social stability,Internet public opinion,and early warning.It focuses on explaining that the main body of Internet public opinion is all Internet users.The characteristics of Internet public opinion are extensiveness,anonymity,and bias.At the same time,it has the functions of information dissemination and publicity orientation,early warning and checks and balances.Secondly,on the research of communication theory related to internet public opinion,the life cycle of internet public opinion is divided into five stages: budding period,development period,peak period,fall period and decline period,and the characteristics of internet public opinion are different in different life cycles.Finally,a method of deductive reasoning and practical experience summary is used to establish a social stability early-warning indicator system,and the literature research method is used to analyze the impact of Internet public opinion on the first-level indicators in the indicator system,which will eventually affect social stability.Therefore,the influence of Internet public opinion on social stability cannot be ignored,and research on early warning of Internet public opinion is of great significance for maintaining social stability.The core content of this article is the establishment of an online public opinion early warning model.First,this article constructs an online public opinion early warning indicator system including 4 primary indicators and 16 secondary indicators,including public opinion enthusiasm,public opinion participation,public opinion themes,and public opinion dissemination trends.,Clarified the specific meaning of 16 secondary indicators and quantified the indicators.Secondly,based on the entropy method-CRITIC weighting method and BP neural network,a network public opinion early-warning model was established,and an empirical analysis was performed with the "Liangshan forest fire incident" as an example.Firstly,the public opinion index data was captured from Baidu and Sina Weibo and preprocessed.The weights were calculated using the entropy method-CRITIC weighting method,and the weights of public opinion subject sensitivity and Weibo comment rate were higher.The impact is greater.Then use BP neural network to build a network public opinion early warning model for early warning.Examination of the test samples shows that the early warning accuracy rate of both models reaches more than 60%,which proves the rationality of the index construction and the feasibility of the two models.Because the BP neural network model has a higher warning accuracy rate in the warning evaluation,the entropy-CRITIC weighting method can obtain the weight of the warning indicator system,and can grasp the influence of different indicators on the warning level.Therefore,in the early warning application,two types of early warning models can be used together according to actual needs.