Evaluation and design of online cooperative feedback mechanisms for reputation management

TLDR

The study evaluates current online reputation systems that use cooperative feedback and proposes a new exponential‑smoothing based design. The authors evaluate existing cooperative‑feedback reputation mechanisms, propose a new exponential‑smoothing design, and implement a two‑level smoothing policy that relaxes static product value assumptions. Simulations reveal that the proposed exponential‑smoothing policy offers sustained incentives for sellers and is more robust than existing feedback systems, which fail to encourage long‑term honesty.

Abstract

This research evaluates and analyzes existing mechanisms of online reputation systems based on cooperative feedbacks of past transaction information. We find existing popular feedback systems do not provide sustained incentives for sellers to behave honestly over time. We propose a new design of reputation system based on exponential smoothing. This mechanism is shown to be more robust compared to the existing systems. We relax the assumption of a static product value and have implemented a two-level exponential smoothing policy to process and aggregate reputation information. The simulation results show that the policy can serve as a sustained incentive mechanism for the seller.