Inspired by the ability of the trust mechanism which can capture human trust using measurement theory we have designed trust-based security mechanism to find clusters of fake users in social networks. Proposed framework enables us to take human and community knowledge into a loop by using a variation of trust to detect fake users. This framework will increase the accuracy of detecting fake users by using both human and machine knowledge. Experiments are performed on randomly generated social networks to validate the potential of this framework. Results show that the variation of trust over time can able to differentiate between clusters of fake users from clusters of real users.
展开▼