Analysis Of The Theme Evolution Of Internet Public Opinion Based On Micro-blog Text

With the growth of Internet users,the steady growth of infrastructure and the popularity of smart phones,the development of information network industry has been greatly promoted.At the same time,people's self-awareness and awareness of social responsibility awakening are increasing in democracy,rule of law,fairness and justice.The real time and interactivity of micro-blog meet the public's demand for information and expression in public emergencies,thus making micro-blog a key position to produce network public opinion and network public opinion.In order to help the relevant departments to understand the progress of emergencies,the development of public opinion and the attitude of the netizens better,so as to improve the processing efficiency and the ability to monitor the public opinion,this paper analyzes the theme evolution trend of the network public opinion through the combination of the word weight and the theme model,and mainly studies the following contents:(1)This paper summarizes the current research situation of micro-blog's research themes,thematic analysis and public opinion analysis,and elaborates on the theories and methods needed in the research.(2)On the basis of previous research,according to the theory of life cycle and the rule of public opinion development of emergencies,network public opinion is divided into the beginning stage,the outbreak stage,the repeated stage and the long tail stage.The repeated stages of Internet public opinion are further divided into the first recession stage,the second outbreak period and the second recession stage.(3)On the basis of the above research,taking the blue Qianjiang arson case as an example,the theme of the Internet public opinion caused by the sudden incident is analyzed.First of all,through the form of keyword search,we collect hot micro-blog related to the blue Qianjiang arson case from micro-blog.The data collected include the blogger ID,the micro-blog text,the time,the number of points,the number of comments,the commentary ID,the comment time,the comment text,and the stages of the public opinion according to the specific emergencies.Then,we use LDA topic model,feature weight and clustering method to construct the theme model of network public opinion evolution.Finally,the text information of micro-blog is excavated to determine the theme evolution process of the network public opinion in different periods.Through the above empirical research,the following conclusions are drawn:(1)the theme feature words can be used to show the trend of unexpected events;(2)the growth and decline of network public opinion are independent of the number of topics,and related to the progress of emergencies;(3)once the different negative themes are too concentrated,it will lead to the outbreak of network public opinion.