[Objective] Implementing microblogging untrusted users’ identification with subjective uncertainty by improved Dempster-Shafer evidence theory.

[Methods] Firstly, the original Dempster-Shafer evidence theory is improved by correcting the evidence distance. Then, in order to generate the users' trust interval, the credibility of historical posts are transformed into evidences and then merged by improved Dempster-Shafer evidence theory. Finally, according to the users' trust interval, untrusted users are identified by means of Decision Tree algorithm.

[Results] Compared with the existing untrusted user identification methods, the time consumption of our new method is reduced by 287.4 seconds, the   score is increased by 31.9% and the Chi-Square value of the consistency test is optimal.

[Limitations] Only the subjective uncertainty caused by time decay and evidence conflict is considered, and the impact of cognitive differences on subjective degree is not considered. [Conclusions] The identification method of untrusted users based on improved Dempster-Shafer evidence theory is more effective.