Improved Trust Model based on Centrality Measures and Recommendation in Social Network

Abstract

In this article, a model is developed to improve trust value in the relations that represents social networks by utilizing centrality measures interred with all participants in network-based and recommendations, both on connection or trust. Various central metrics were discussed and implemented to intend to trust. Algorithms are provided to facilitate their calculations. The referral neighbor that has a guaranteed trust boundary is chosen. Trust Value based on Interaction and Recommendations using Centrality Metric (TVIRCM) method is proposed and implemented in this study to improve trust value in the social network when the link between any two nodes represent the unique indication about trust whereas, there are no other standards for maintaining trust. Trust based on interaction refers to trust calculations based on real links observations and exploits the centrality metric. Trust based on recommendation refers to trust calculation based on trust participants of a remote neighbor about other participants.The developed approach is utilized in a trust observation phase as a trust-based interaction (i.e. assigned high and low centrality metrics), then the next phase is based on the proposed recommendation (i.e. the remote neighbor may have certain trust value) and the last phase is the trust calculation phase which based on combining direct and indirect trust.