Perseus - A Personalized Reputation System
In: The 2007 IEEE/WIC/ACM International Conference on Web Intelligence, 2-5 Nov 2007, Fremont, California.
We propose Perseus, a personalized reputation system. In Perseus, reputations comprise of three aspects: how much I personally trust another individual, how trustworthy others think the individual is, and how much I trust the opinions of others. Perseus is adaptive in the sense that user feedback is used to modify the way the different aspects are considered. We also present simulation experiments, which indicate that Perseus is robust and able to survive under extreme conditions of misbehavior. In addition, Perseus encourages individuals to rate the other party and give fair ratings. We also compare Perseus against other well-known reputation systems.