A Trust-Based Reputation Management System

Since its inception in the early 1990s, e-commerce in consumer-to-consumer (C2C) markets has \nachieved great success, with significant projected growth. For example, the Internet auction provider, \neBay, has established itself as the largest global player in this market, with $34.2 billion worth of \nmerchandise being auctioned in 2004 and 135 million registered users in 32 markets worldwide. The \nC2C domain, analogous to its conventional physical marketplace equivalent, is built on trust. Buyers \nsend payments to complete strangers from whom they have purchased goods and trust that the goods \nwill be sent in return. Sellers trust buyers to make good on their payments. All users risk loss, both \nfinancial and of their time. Users establish reputations about their trustworthiness through an \nintegrated feedback collection and distribution system, i.e., a reputation management system. Thus, \nan online marketplace approximates its traditional predecessors as a system in which the human \nconcepts of trust, risk, and reputation are critical to performance. \nThe apparent benefits of interacting in such a strongly-networked global market are accompanied by \ninnovative adaptations of traditional hazards. The Internet, while connecting disparate user groups to \nincrease transaction potential and shared knowledge about the marketplace, also permits user \nanonymity and transactional intangibility, which can lead to fraud, theft, and collusion. Reputation \nmanagement systems attempt to limit incorrect behaviour and to assist decision making by providing \nrecords of feedback about interactions, called recommendations, for each community participant. \nThese systems are not without their own limitations. First, commercial reputation management \nsystems typically promote usability over accurate evidentiary analysis, meaning that data which could \nbe extremely useful to decision-making is disregarded by the evidence collection mechanism so that \nease-of-use is maintained for community members when they are voluntarily providing feedback. \nThis first issue leads directly to the second, which is inaccurate evidentiary analysis with regard to \ncontextual relevance in terms of user role, timeliness of evidence, and environmental context. In this \nregard, trustworthiness is usually linked solely to the overall number of positive recommendations \nabout a user, regardless of the interaction context being considered. Third, the dynamics of user \ninteractions are not addressed, and interaction dynamics in such an evidence-rich environment are \ndifficult, if not impossible, for an average user to manually detect. Without the ability to analyse \ninteraction dynamics, the fourth and fifth issues arise, namely that the analysis of whether or not a \nuser provides useful and accurate recommendations about another user or whether or not a group of \nusers are colluding with malicious intent are both difficult to observe. Sixth, risk is not explicitly \ncalculated by the reputation system, and may not be assessed by the user at all. Seventh, and finally, a \nreputation is often no more than an overall summary of a collection of thousands of individual \nrecommendations rather than an explicit portrayal of the trust and risk involved in a context-specific \ninteraction.