A Research On Recognizing And Studying Negative Financial Public Opinion

In many areas of Internet public opinion,financial public opinion is an important issue of people's livelihood,which reflects the real interests of people.With the development of network media,Internet users' views on financial events can be transmitted rapidly through the network.Some research work has showed that Internet users' financial sentiment tendencies,stocks and market volatility are closely linked,which can raise the risk of the lawlessness of the lawless man.Thus,studying on monitoring financial public opinion and sentiment tendency has important significance to master the financial market,to sense the national economic micro-environment and maintain social stability.For the financial public opinion analysis,we believe the study on negative information has important research value.Negative information can cause the bad emotion of recipients,and even lead to the emotional drama in a bad situation,which can cause the herding effect,and thus greatly damage the reputation of an institution and then affect the financial markets.Although the study of text tendency has been used widely in the field of public opinion analysis,the existing research is still lack of in-depth discussion of negative information recognition.In view of the blind spot of the current research,this article takes the negative information recognition of financial public opinion as the research topic,selects the net loan industry which has great influence on social stability in 2016 as the research object,crawls the full-year net loan industry public opinion data from Sina Microblog as data sample.Based on the establishment of the negative information recognition model framework,and marking a large number of network loan public opinion text data,this article makes the in-depth analysis of financial public opinion from the lexical,syntactic and semantic level.The research method combines the new word discovery,the short text feature extension based on the deep learning,the feature weight criterion discussion,the text classification,the negative index,the negative subject analysis,and the event extraction based on the dependent syntax.The research results can reflect the current situation of the negative information of the network loan industry from the macro and micro perspective,and provide the basis for the future research of negative information and reference for the triggering words or indicators of financial negative information monitoring.