Microblog Credibility Assessment Based on Multi-dimensional Method

摘要

Information is shared and disseminated in the microblogging platform, but not all the information is trustworthy. In the research, the microblog credibility assessment is mainly about the statistical analysis of the static feature data distribution at a time point, not considered the estimators subjectivity and trust dynamic. For this purpose, Microblog Credibility Assessment Based on Multi-dimensional Method (MCMD) is proposed, this method considers the subjectivity and dynamic effects of trust. From user, content and time dimension to assess the credibility respectively. By introducing the weights system to satisfy the subjective preference of estimator, this method supplement objective basis, and dynamic adjust the model based on nonlinear time factor. Experiment results show that this method reflects the influence of subjectivity and dynamic on the credibility assessment, and are significantly more accuracy than the traditional classifier approach.