Locating Experts via Online Social Networks

Online social networking systems provide indirect access to a large number of people connected by multi-step chains of acquaintances, and plays an important role in the referrals for human information flow. In this paper, from a networking point of view, we study the problem of locating experts for relevant information via online social networks. We model the action of forwarding a question with random walk, adjusted by a node´s awareness of the potential expertise of his immediate neighbours. Using the model we derive analytical expressions of the performance metrics of a referral session in terms of the nodes´ awareness level of their neighbours and the percentage of nodes that may have answers to the posed question. We also utilize several real online social networks to study the modeled question-forwarding strategy, and find that the simulation results validate our analyses.