A Demspter-Shafer approach to provenance-aware trust assessment

In this paper we present an agent-based approach to managing the trustworthiness of information in network centric information sharing environments. Specifically, we focus on how to avoid the double counting problem in the process of trust assessment in a principled way. We describe how provenance information can be used to detect and avoid double counting during trust assessment processes. In our approach, each operator is associated with a software agent, called client agent. The client agent collaborates with other agents to automatically assess and learn the trustworthiness of information from its provenance within the framework of Dempster-Shafer theory. We empirically evaluate the effectiveness of our approach and show how users´ feedback can be captured to enhance trust assessment in a simulated information sharing environment.