Trust Evalution Research Of Sina Weibo Users And Information

With the popularity and development of Online Social Network(OSN),Social Network plays a more and more important role in people's daily communication.People can publish,obtain and transmit information in the social network,interact with other users and build their own social network,and make decisions based on the information obtained in the social network.However,there are a lot of untrue and misleading information,as well as untrustworthy users in social networks.Therefore,it is of great significance to establish the trust evaluation mechanism of users and information.At present,most of the studies on evaluating users and blogs in social networks take foreign social network platforms such as Twitter and Facebook as the research objects.As one of the largest Chinese social network platforms in the world,Sina Weibo is significantly different from foreign social networks in terms of scale,user groups,publishing settings and user communication modes.This paper first compares the differences between Sina Weibo and foreign mainstream social networks,on this basis,proposes a trust evaluation method for Sina Weibo users and blog posts,and designs and implements a trust evaluation application system for Sina Weibo users and blogs.The main contributions of this paper are as follows:(1)Analyzing the semantics of the trust in online social networking platform.On the basis of Co Rank algorithm,for the environment of Sina Weibo,introducing the relationship of "thumb up",improving the algorithm,defining the "composite actions with the thumb up" and other complex scenarios.These scenarios are discussed in a social network model represented by the semantic and the corresponding weighting method.For the problems of weight filling conflict,rank sink and data missing,solutions are put forward.Finally,SWCORANK(Sina Weibo CORANK)algorithm is proposed.(2)A web crawler was written to collect the data of users and blogs in Sina Weibo,and based on the collected real data,the trust evaluation was conducted on the collected users and blogs.The performance of the proposed SWCORANK was verified through experiments.In the analysis of experimental results,on the one hand,the current mainstream methods such as Page Rank and TURank are used to calculate the same data set and compare the results with each other to prove the effectiveness of SWCORANK.On the other hand,sample users and blogs are analyzed and explained through sampling to visually display the algorithm evaluation results.(3)The "Sina Weibo trust evaluation system" is designed and implemented based on Flask framework.For the uploaded network structured data,the system calls SWCORANK(Sina Weibo CORANK)to generate the corresponding trust ranking.The system not only visualized the social network structure at a given moment,the social relationship of a given Weibo user and blog,trust ranking,etc.,but also visualized the dynamic change information of the above information as the social network changes over time.At the same time,it is convenient to see the influence of interaction and social relations such as following,fans on trust ranking.