A novel reputation-based model for e-commerce

Abstract

Reputation systems are very important in e-markets, where they help buyers to decide whether to purchase a product. Since a higher reputation often represents higher profit, malicious users may try to deceive them to increase their reputation. This may be aggravated through fear of retaliation to bad ratings or collusion of multiple users so that they can perform a fraud of high value before being denounced by their victims. In this paper we provide a new reputation management system (RMS) which prevents retaliation and minimizes fake ratings due to collusion. Furthermore, it combines the price and the sales track record of the seller with the charges he has already paid to the system. In this way, it becomes too expensive both in terms of time and money for the malicious users who may collude to launch an organized attack discrediting an e-commerce platform. We test our proposals with several simulation scenarios, verifying the validity of our claims.