In online social networks, reputations of users (nodes) are emerged and propagated through interactions among the users. These interactions include intrinsic and extrinsic consensus (voting) among neighboring users influenced by the network topology. We introduce an algorithm that considers the degree information of nodes (users) to model how reputations spread within the network. In our algorithm, each node updates reputations about its neighbors by considering the history of interactions and the frequency of the interactions in recent history. The algorithm also captures the phenomena of accuracy of reputations deteriorating over time if interactions have not occurred recently. We present the following two contributions through experiments: (1) We show that an agent’s reputation value is influenced by the position of the node in the network and the neighboring topology; and (2) We also show that our algorithm can compute more accurate reputations than existing algorithms especially when the topological information matters. The experiments are conducted in random social networks and Autonomous Systems Network of the Internet. In addition, we show the efficacies of each component in our algorithm and present their effects on the algorithm.
https://www.researchgate.net/publication/295240343_A_Node-Centric_Reputation_Computation_Algorithm_on_Online_Social_Networks
In online social networks, reputations of users (nodes) are emerged and propagated through interactions among the users. These interactions include intrinsic and extrinsic consensus (voting) among neighboring users influenced by the network topology. We introduce an algorithm that considers the degree information of nodes (users) to model how reputations spread within the network. In our algorithm, each node updates reputations about its neighbors by considering the history of interactions and the frequency of the interactions in recent history. The algorithm also captures the phenomena of accuracy of reputations deteriorating over time if interactions have not occurred recently. We present the following two contributions through experiments: (1) We show that an agent’s reputation value is influenced by the position of the node in the network and the neighboring topology; and (2) We also show that our algorithm can compute more accurate reputations than existing algorithms especially when the topological information matters. The experiments are conducted in random social networks and Autonomous Systems Network of the Internet. In addition, we show the efficacies of each component in our algorithm and present their effects on the algorithm.