Bayesian game theory in practice: A framework for online reputation systems

As the number of online auction sites has increased the interest towards providing reliable summaries, reputations, about the past behaviour of the sellers has risen. However, many of the existing approaches are based on various heuristics and properly evaluating the systems is difficult. To improve the situation, we first introduce a game-theoretical model for reputation in online auctions. Furthermore, we show how mixture models can be used as a proper framework for integrating information from various sources into the model. We also discuss practical aspects and evaluate an experimental online reputation system that has been implemented based on the proposed framework. Our results show that, using this framework, the number of successful transactions increases significantly (in statistical sense) and that the framework is robust for changes in the behaviour of the participants. Computing Reviews (1998)