Sensitivity Analysis of Multi-State Social Network System Based on MDD Method

With the advancement of computer and network technologies, Internet-based social networks called social networking services have become popular. Trust is a crucial basis for interactions among parties in social networks. Based on trust scores of direct links between parties, a trust sensitivity analysis can help identify which direct link(s) in a social network contributes the most to a trust relationship between parties who are not directly connected in the network. This paper generalizes the research object from two-state social networks to multistate social networks since the trust grade for people in a real social connection may have multiple levels. We model asymmetric multitrust level and multiparty social network systems and propose a probabilistic method based on multivalued decision diagrams (MDDs) to perform trust sensitivity analysis of social networks. Numerical examples are provided to demonstrate the application of the proposed methodology.