Font Size: a A A

Analysis And Prediction Of Group Behavior Of Network Public Opinion Based On Rumor And Anti-rumor

Posted on:2021-04-10Degree:MasterType:Thesis
Country:ChinaCandidate:Z HuangFull Text:PDF
GTID:2428330614458391Subject:Computer Science and Technology
Abstract/Summary:PDF Full Text Request
Since the Internet came into China in the 1990 s,mobile terminal technology has been rapidly developed and widely used.Online social network has gradually replaced traditional media and become the main carrier of people's information dissemination with its characteristics of super space-time,openness and virtuality.Social network provides great convenience for people's life,but also provides a channel for rumors.Compared with traditional media,online rumors are more likely to breed,spread more widely,and spread faster.Therefore,it is very important to predict the spreading trend of rumor effectively for restraining rumor spreading and building a healthy network environment.Based on the content,structure and behavior characteristics of the rumor and anti-rumor topic communication space,this thesis models,analyzes and studies the topic situation change at the individual level and the group level.The main research work and contributions of this thesis are as follows:1.At the individual level,a rumor and anti-rumor information oriented user behavior prediction model is proposed.Firstly,aiming at the privacy of user relationship,a user2 pixel algorithm is designed to mine the privacy of user relationship,and the topic propagation user relationship network is transformed into image pixel array.Secondly,aiming at the antagonism of rumor and anti-rumor information,considering the internal and external factors that affect users' behavior in the process of rumor propagation,a quantitative rumor and anti-rumor mutual influence model is established.Finally,based on the construction of topic propagation pixel array by user2 pixel,considering the advantages of convolutional neural network in simple processing of topic propagation pixel array,and referring to the rumor mutual influence transfer matrix,a rumor and anti-rumor information oriented user behavior prediction model is proposed based on convolutional neural network.2.At the group level,this thesis proposes a topic heat prediction model for rumor and anti-rumor information.Firstly,in view of the difference between the user's activity and the user's relationship in the topic communication space,the user's communication power and user interaction degree are introduced,and a double weighted user relationship network is constructed.Then,based on Page Rank algorithm,an algorithm w-pagerank is designed to quantify user influence.Finally,in view of the influence of rumors and anti-rumor information on the popularity of rumors,we design a prediction model based on the influence of users and anti-rumor information by using regression algorithm and combining the characteristics of early topics and user influence.Finally,through the data set of sina Weibo published by Tsinghua University team to verify.The experimental results show that the model and algorithm proposed in this thesis can effectively predict the individual communication behavior and group communication trend in the process of rumor topic communication,and can timely find the main users promoting rumor communication.Therefore,this study plays an important role in the guidance and control of Internet public opinion and the construction of a healthy Internet environment.
Keywords/Search Tags:rumor&anti-rumor, dissemination pixelation, representation learning, evolutionary game, group behavior
PDF Full Text Request
Related items