Public Sentiment Collecting And Tendencies Analysis Based On Sina Microblog

With the rapid development of the Internet, penetration of the Internet is higher and higher, people are increasingly willing to express personal views on the social hot spots and emergency through the network. Today’s online public opinion subvert the traditional network model, it starts with an unstoppable trend into the social dimension, resulting in a very profound impact on society. Microblogging has powerful interactional function, highlighted instantaneity, convenient phone support, celebrity and many other advantages, it has been gradually replaced blog, post bar and other veteran public opinion communication platform, and it feed into the development of online public opinion, it has now become the most important network communication platform of public opinion.Microblogging attention function can associate users together so that speed of information dissemination has been changed historically, which also makes it possible to use microblogging information to study and master the development direction of public opinion, and provide scientific analysis results for government and business, in order to enhance their decision-making capacity, this makes high social value and commercial value.The system consists of public opinion collection module, public opinion pretreatment module, public opinion analysis module, public opinion service module and system management module. We select Weibo as a data source for public opinion analysis system, and use java timing tasks to get content form Weibo regularly, decompose obtained content pages according to the structural characteristics of Sina Weibo page, store in the database, then use clustering, vectorization and other classification means to analysis, finally we give results using orientation analysis, and show the users the trend of public opinion over time in the form of charts based on analysis of the public opinion. On the software level, the system select SpringMVC, Spring, Hibernate and other open source technologies as the whole architecture,in order to reduce the degree of coupling between the modules and improve program scalability. in the acquisition and analysis module, we also use MySQL, HtmlParser, ICTCLAS Chinese word system and other open source technologies.In the process of developing system,we use unit testing, functional testing, performance testing and security testing.During testing,we chose Monster Hunt as our keyword to acquire data,ultimately,we get 432361 related microblogging between July 9,2015 to July 21,2015.By duplicate removal, de-noising, semantic analysis, clustering, classification of speech and a series of operations,we eventually acquire public opinion trends about the topic of Monster Hunt. We chose movie release date July 16 as a watershed, before release,the percentage of positive microblogging related to Monster Hunt is 45%, and increased to 90% after the release. In practice, the film side can adopt different promotional strategies based on characteristics of public opinion in different periods, in order to maximize benefits.