Reputation Systems: A framework for attacks and frauds classification

Abstract

Reputation and recommending systems have been widely used in e-commerce, as well as online collaborative networks, P2P networks and many other contexts, in order to provide trust to the participants involved in the online interaction. Based on a reputation score, the e-commerce user feels a sense of security, leading the person to trust or not when buying or selling. However, these systems may give the user a false sense of security due to their gaps. This article discusses the limitations of the current reputation systems in terms of models to determine the reputation score of the users. We intend to contribute to the knowledge in this field by providing a systematic overview of the main types of attack and fraud found in those systems, proposing a novel framework of classification based on a matrix of attributes. We believe such a framework could help analyse new types of attacks and fraud. Our work was based on a systematic literature review methodology.