Font Size: a A A

Research On Information Credibility Evaluation Method For Social Network Information Control

Posted on:2018-09-23Degree:MasterType:Thesis
Country:ChinaCandidate:J Q LuFull Text:PDF
GTID:2348330563451359Subject:Systems Engineering
Abstract/Summary:PDF Full Text Request
With the development of Internet technology,WeChat,QQ,weibo,Facebook,Twitter and other social network platforms with their ease of use,real-time performance and convenience have become important channels for people to obtain information.But they also provide a convenient platform for the dissemination of false information and malicious rumors.So how to effectively control the social network information and the spread of false information and rumors is of great significance to maintain social stability and ensure the sound development of the social network.This paper can control the release of social network information through information filtering,information credibility evaluation,public sentiment monitoring and other technical means.However,due to the diversity of social network structure,the complexity of the information content and the uncontrollable information publisher,it brings great challenges to information management.According to the characteristics of the release of social network information,we combine the multilevel evaluation of network information content credibility and the dynamic assessment of user credibility organically,and control the release of social network information according to the strategy.The research emphasis mainly includes the following several aspects.(1)On the basis of analyzing the characteristics of social network information release and control,we propose a framework of social network information management and control based on credibility evaluation which is available to managers of social network information release.We also analyze and design the process flow of multilevel evaluation and control strategy of social network information content credibility and the dynamic assessment of user credibility and control strategy.(2)According to the characteristics of Chinese text in social network information content,we propose an improved algorithm which is based on Bayesian algorithm classifying the social network information content credibility,and address the problem which has not considered the semantic relevance in traditional Bayesian classification algorithm processing text.The experiment results show that the improved algorithm is more suitable for Chinese semantic expression and improves in processing accuracy.(3)This paper propose a text similarity calculation method that can detect rumors rapidly.For the information which is unable to be determined,we put forward an information content credibility evaluation method based on analytic hierarchy process(AHP)through extracting the number of URLs and the number of emotion words,and adopt machine learning to evaluate the of information content credibility dynamically.(4)This paper propose a user credibility dynamic evaluation algorithm,which adopts probability method to evaluate user reputation based on users' built-in attributes.On this foundation,we combine the credibility of information content that users release,and realize the dynamic adjustment of user credibility.The experiment results show that the proposed algorithm has a significant improvement in processing accuracy.
Keywords/Search Tags:Social network, Information control, Information credibility, User credibility, Evaluation method
PDF Full Text Request
Related items