Research On Evolution Of Network Public Opinion Theme Events Based On Social Network Analysis

Along with the development of information technology,cyberspace has become an important place for the populace express an opinion or attitude about an event,thus forming a network public opinion.Network public opinion,through the Internet,has the characteristics of real-time information,audience universality,beyond time and space,spread quickly.It is a reflection of the public opinion events in the real world,making the events spread quickly using its particular features.These new characterist-ics raises new challenges to timeliness and accuracy of the government governs public opinion events.Therefore,strengthening the research of network public opinion,especially network public opinion evolution,has an important value and practical significance.The existing research mainly studies the evolution of network public opinion from two perspectives.First,from the view of user's participation in network public opinion,dividing the number and the characteristics of participated in the network public opinion into stages,or using model simulation to study the users'change of participate in public opinion events according to the mode of public opinion transmission.Second,from the view of public opinion texts,studying sentiment analysis,theme discovery and the evolution of theme popularity.However,it is impossible to systematically reveal the evolution activity of network public opinion theme event from user's perspective or from the view of network public text respectively.Therefore,based on the above two research ideas,this paper proposes a social network analysis method to carry on a comprehensive research on users participated in network public opinion events and its text contents.Then,using text mining technique and system simulation technology to systematic study the evolutionary trend of network public opinion theme events.This paper first reviews the research status about network public opinion,the evolution of network public opinion,social networks,short text mining.Secondly,it introduces the theoretical basis,including network public opinion,social network analysis,text mining and system simulation.Then,it builds user networks and text content networks of network public opinion theme events and studies the overall network structure based on the actual public opinion event data.On this basis,it analyzes the content structure of network public opinion theme,analyzes and verifies the characteristic indexes of social network,explores the evolution mechanism of network public opinion.Finally,based on the whole user networks,using the improved model SEIR as the propagation evolution model,it verifies the key factors of influence the evolution of public opinion events,and proposes the corresponding public opinion strategy.The results demonstrate that affecting the evolution of network public opinion theme events is related to certain key factors including the number of users and the status of users participating in the propagation event,also the theme interest index and attention hot theme of users participating in the event.Therefore,this paper proposes that the governments not only strengthen the guidance and control of key users participated in network public opinion events.They also need to detect theme focus on their attention to public opinion events,then adequately disclose relevant public opinion information,evacuate public opinion blind spots,guide and control the trends of public opinion events.