Research On Enterprise Public Opinion Monitor Method Based On Sentiment Analysis

With the rapid development of Internet industry,the internet opinion,which was regarded as a manifestation of public opinion,has brought a lot of topics to enterprises and hidden a huge risk at the same time.Negative comments from netizens and negative press coverage by the media are likely to result in the loss of consumers or stop the investors.For each business manager,it's very necessary to have a clear grasp of the business assessment of the situation,to understand their own business network of public opinion risks on the network.However,most of our Internet public opinion monitoring network are aim to illegal content such as crimes and pornography.Although some of the internet public opinion monitoring has achieved good results,these kind of monitoring is often used by government agencies.For enterprises,the monitoring content is less relevant to the enterprise,while the average user and media sentiment Attitude should be the key factors.Based on the analysis of the network public opinion monitoring research results at home and abroad,this paper studied the core technologies of public online monitoring from the perspective of textual emotion analysis,improved the texts and emotions Lexical matching degree by extending the sentiment dictionary of the University of Taiwan and the HowNet sentiment dictionary,achieved the classification of text neutral emotion through the establishment of a neutral word dictionary,finished the construction of the degree of adverbs and other features dictionary to extract textual affective features,combined with naive Bayesian model and decision tree model,three categories of emotional polarity have been completed for online commentary data and online news data,improve the PMI algorithm and enhance accuracy by extend standard word set,based on the data of emotion analysis,the paper developed the corporate internet public opinion monitoring indicators,by combing with improved fuzzy comprehensive evaluation and analytic hierarchy process,the paper also offered the corporate network public opinion monitoring methods.Finally,the case analysis was carried out based on the network public opinion information of 9 enterprises.After a lot of theoretical research,this paper designs and implements the network public opinion monitoring system based on the sentiment analysis,the system includes the web crawler module,the data storage module,the sentiment analysis module and the data presentation module.Through the system,business managers can have a more intuitive understanding and real-time monitoring of their own public opinion crisis on network;ordinary users can also view the corporate public opinion as an important reference to the consumer,transaction,employment and other decisions.