Public Opinion Analysis Model Of Trade War Considering Policy Factors

With the popularity and rapid development of the Internet,the network public opinion has gradually become a very important force for social public opinion,and to a large extent,it guides the Internet users to understand certain public opinion events.Because of the characteristics of rapid dissemination and large amount of information,how to quickly discover and analyze network public opinion information has become a hot research topic.As one of the focus issues,trade war has aroused great concern of the general public,and has rapidly formed into a huge network of public opinion on the Internet.Trade warfare based on policy considerations is the direction of public concern.Compared with other network public opinion information,the trade war considering policy factors embodies more policy contest between some governments,which is closely related to the interests of the masses and the comprehensive strength of the country.This paper adopts topic models to discover and analyze public opinion information of trade wars based on policy considerations.And it analyses the changing trend of some major incidents in trade wars and the results of policy factors,including "Tariff incident","ZTE incident" and "Reconciliation incident".On this basis,the extracted feature words and their weights are analyzed,and the specific work is as follows:(1)Collection of text data based on public opinion information of trade warfare considering policy factors: This paper adopts the basis of Web crawler technology in the theme,set up a set of topics such as trade war,gold,stock market,import and export,economy and so on to crawl relevant text data on microblog.(2)Text preprocessing: LDA topic model can only deal with structured text data,so,the crawled natural language is processed.Firstly,we use the combination probability of each word to count and segment the text.Then we use the TF-IDF model to process the data structurally.It weighs the weight of the feature words between a single document and all document sets.(3)Using LDA topic model to find the feature words we need and get the feature words we need and its weight in the document.(4)The experimental results are analyzed and introduced in detail.Based on the above work,this paper comes to the conclusion that analyzed public opinion information on trade war between enterprises and government from a macro perspective.The similarities and differences between them are analyzed."ZTE incident",as an enterprise’s trade war public opinion information,and "Tariff incident" and "Reconciliation incident",as a government’s trade war public opinion information,all obey the development cycle of public opinion information,and the introduction of policies will affect the emergence of the peak.The trade war public opinion information related to the government will have multiple peaks,and the second peak will have the policy superposition effect,and the cycle is long.Areanalyzed from the aspects of micro trade war based on the policy factors underlying theme of public opinion,this article got six implied themes: the government,the domestic attitude,gold,import and export tariffs,the stock market and economy,the government and domestic attitude for Public and Government Attitudes orientation implied theme,gold,import and export tariffs,the stock market and economy for the economic aspects of the underlying theme.Furthermore,the underlying themes are analyzed in detail with the obtained characteristics and weights.At last,the conclusion is used to give some suggestions on the guiding time from the development cycle of microblog.Suggestions are given for the content based on the implied theme.