Prescriber-Consumer Social Network Analysis for Risk Level Re-estimation based on an Asymmetrical Rating Exchange Model

In this paper, we present a novel approach to reestimate the risk level of prescribers and consumers (doctors and patients) that were previously evaluated by various independent Risk Analysis Systems (RAS). This is achieved by taking into consideration social network structure between prescribers and consumers. A mathematical model, called Asymmetrical Rating Exchange Model (AREM) is proposed to describe the mutual influences between prescribers and consumers from a social network perspective and based on this model an algorithm is derived to reestimate the suspicion level of each entity in the community considering both the pre-evaluated rating and the network structure. Additionally, by comparing the pre-evaluated rating and the re-estimated risk level, under-rated entities can be identified and further assessed. Experimental results are also presented, showing that the proposed approach can effectively detect the entities that have strong connections to high risk entities, but previously been rated low suspicious by independent Risk Analysis Systems (RAS).