Research And Implementation Of Micro-blog Public Opinion Visualization System

With the rapid development of Internet technology and the continuous in-depth implementation of the "broadband China" strategy, the network media has become the main channel of information dissemination, and its influence on the social public opinion has also become increasingly significant. Along with the convenience it brings, micro-blog is becoming a hotbed for fake information breeding and flooding. Therefore, analysis and monitoring on micro-blog public opinion is of great practical significance.With the explosive growth of micro-blog users, and with the exponential increase of micro-blog information, the traditional presentation of micro-blog public opinion analysis result has been unable to meet the actual needs gradually. From the perspective of presenting public opinion, this paper analyzed the current relevant research of micro-blog opinion visualization. And on the basis of the study of the graphical presentation of micro-blog public opinion, this paper came up with a visual method of presenting current micro-blog public opinion general situation. Then, based on this method, a micro-blog public opinion visualization prototype system was designed and implemented.The main work of this paper are as follows:At first, according to the analysis and study of public opinion visualization, this paper came up with a visual method of presenting micro-blog public opinion general situation and finally implemented it by using D3.js library. This visual method intuitively presents the data of micro-blog hot topics, their attention degree, users sentiment and the potential associations among them.Secondly, in the process of micro-blog pretreatment, Chinese word segmentation and Chinese POS tagging was completed by calling the API of NLPIR system, and then through POS, length, context collaborative filtering for word segmentation results, VSM representation of micro-blog text was obtained.Thirdly, in the process of micro-blog public opinion analysis, semantic similarity of micro-blog texts was calculated based on HowNet knowledge system, and the semantic distance between micro-blog texts was measured, and then through the application of Single-Pass Algorithm, hot topic of micro-blog was detected. Also, in this process, Micro-blog hot topic sentiment was analyzed by using a sentiment lexicon built on beta version of HowNet sentiment words set.At last, based on the study of key content in public opinion analysis, a micro-blog public opinion visualization system was analysis and designed in detail. And then a prototype micro-blog public opinion visualization system was implemented considering the visual method of presenting micro-blog public opinion general situation. And then, considering the proposed visualization method, a prototype of micro-blog public opinion visualization system was implemented. This prototype visualized the general situation of public opinion by analyzing the input micro-blog texts data set.