REAL: A risk-enabled reputation model for electronic commerce

Recently, a trend to use reputation model for selecting service providers becomes evident in the area of electronic commerce. Most of the effort have been made on providing a reliability value associated with the predicted reputation to interpret how reliable is the reputation. However, the predicted reputation will become ineffective with a low reliability value. Rather than providing a reliability of a reputation, in this paper, we propose a reputation model, called Risk-Enabled reputAtion modeL (REAL), by introducing the notion of risks to better interpret the recent performance of a service provider in e-commerce. The proposed approach has been validated as a better model to reflect a service provider's recent performance than given by the extant on-line auctions sites or the models using the technique of adopting mean values.