Public Opinion Analysis System For Micro-blog Hot Events

Because of the popularization of Internet,a great deal of news and information related to social hot events spread through the network.The netizen's demands for obtaining information is getting greater.and greater.Since the popularity of the micro-blog and the fact that the netizen participate in the discussion of hot events more widely via the micro-blog,the micro-blog has become an important platform for the netizen to express their views on hot events.On the other hand,with the development of text mining technology,network public opinion analysis based on text mining has a very good feasibility when the platform of public opinion is changing from newspaper and traditional TV media to text information on the network.In addition,because it's hard for the netizen to have a comprehensive and objective understanding of the public opinion of hot events due to massive internet information when they participate in the discussion of hot events,so the netizen's demand for public opinion information is increasing.Thus,this paper brings about the public opinion analysis system for micro-blog hot events.This system collects information of micro-blog and comments related to designated micro-blog events,and uses technologies including information extraction,view identification,sentiment analysis and natural language processing to fetch public opinion information that users concern intelligently,then makes quantitative analysis via statistical method and presents to users via visual interaction.The public opinion system mentioned in this paper includes four modules:the first one is the information acquisition module,which is used to fetch massive micro-blog information and store the information in database for subsequent analysis.The second one is the text preprocessing module,which mainly processes the raw data as a structured data that can be used for public opinion analysis after finishing word segmentation,keywords extraction,named entity recognition,high frequency phrase extraction and public opinion element extraction by using natural language and information extraction technologies.The third one is the public opinion analysis module,which is used to do dimensional analysis including thesaurus construction,public opinion target recognition,hot events elements recognition and situation analysis.The last one is the WEB application layer public opinion information display,which can make users have a more intuitive and comprehensive understanding of hot events public opinion by displaying public opinion information via interactive visualization based on users'demand for public opinion information.While this paper makes great effort on the accuracy,usability and extendibility of this system,and this system realizes designated demands,there are still many places to be improved and studied in the function and performance.