Research On Microblog User And Content Credibility Evaluation Algorithm

Currently,people's access to information is becoming more diverse,from traditional paper media to news apps to the current online social network(OSN).The Weibo platform is one of the OSN.With the popularity of Weibo,people's information from Weibo can't be credible.Because everyone can be the source of news,everyone can become a communicator of news.At the same time,malicious users often publish or spread some false statements in order to achieve their own goals.These statements can interfere with the judgment of normal users and thus be exploited.Therefore,credibility assessment for Weibo is a practical way to filter malicious or irrelevant information for users,purify topic squares and user timelines.The research in this paper is to evaluate the credibility of users in Weibo and Weibo content under specific topics.Aiming at the problem of Weibo user credibility evaluation,this paper proposes a user credibility evaluation algorithm based on trust propagation.This paper considers the highconsumption and low-precision situation of malicious users attacking algorithms and manually selecting seed sets by establishing false social relationships.Three optimization strategies are proposed: pruning algorithm based on social activity and similarity,cluster-based Seed node selection algorithm and propagation algorithm based on two-way trust propagation.The pruning algorithm can trim off the attack edges established by malicious users and normal users.The seed node selection algorithm combines the clustering algorithm with the manual selection to semi-automatically select a set of highly available seed nodes.Finally,the two-way trust propagation-based propagation algorithm uses the user's social relationship diagram to propagate the trust scores in both directions,so that high-confidence users have high-reliability scores,and vice versa.Finally,the effectiveness of the trustworthiness-based user credibility evaluation algorithm in the evaluation of Weibo user credibility is verified by experiments.Aiming at the problem of microblog content credibility evaluation,this paper proposes a microblog content credibility evaluation algorithm based on heterogeneous network.The algorithm is divided into two parts: the similarity-based Weibo credibility score bias initialization and the hybrid propagation rule-based Weibo credibility evaluation.Comparing the verified true news reports and malicious sentiment to different initial credibility scores of Weibo,the microblog credibility score is initialized based on the similarity.In the microblog credibility evaluation of hybrid communication rules,the impact of Weibo content,users and context on the process of trusted score propagation is considered,and the scores are spread between different entities.Finally,it is verified by experiments that the microblog content credibility evaluation algorithm based on heterogeneous network can provide a high-quality trusted microblog data set by iteratively calculating the microblog content credibility score.