Using structural information for distributed recommendation in a social network

Social networks are social structures that depict relational structure of different entities. The most important entities are usually located in strategic locations within the network. Users from such positions play important roles in spreading the information. The purpose of this research is to make a connection between, information related to structural positions of entities and individuals advice selection criteria in a friendship or trust network. We explore a technique used to consider both frequency of interactions and social influence of the users. We show, in our model, that individual positions within a network structure can be treated as a useful source of information in a recommendation exchange process. We then implement our model as a trust-based exchange mechanism in NetLogo, which is a multi-agent programmable modeling environment. The experimental results demonstrate that structural position of entities can indeed retain significant information about the whole network. Utilizing social influence of entities leads to an increase in overall utility of the system.