Evaluating a New Reputation Algorithm With Consideration For Fairness

Abstract

The fairness of reputation and trust management systems is rarely taken into account in their evaluation due to the prevalence of the paradigm of the Prisoner's Dilemma. Yet, considering fairness explicitly leads to a different design of reputation algorithms. This paper proves this claim by considering fairness in the design of a new reputation algorithm that uses the concept of implicit negative feedback: the interpretation of missing feedback as a semi-negative feedback. The use of fairness in the evaluation of the new algorithm shows the opportunities and threats of using implicit feedback.