A Reputation Model Considering Buyer Subjectivity and Product Return Behavior

摘要

With the popularity of e-marketplace, the reputation system has received great attention. For accurately evaluating the trustworthiness of sellers in their transactions, a new computational reputation model is proposed, which considers the buyer credibility. The credibility of a buyer is further determined by the buyer's subjectivity and his/her behavior in returning products. The buyer subjectivity in our paper not only reflects the difference between the buyer's rating and the average rating, but also reflects the fluctuation of the ratings provided by the buyer himself/herself. The product return behavior can directly indicate whether the buyer is satisfied with the goods they purchased or not. It has been shown that the profit-maximized equilibrium in the proposed system can be achieved, where sellers behave honestly in providing products and buyers not return the received products. Theoretical analysis and simulation based experiments have been conducted to demonstrate that the proposed reputation model can well evaluate the sellers' behavior and the proposed system is robust against typical attacking behaviors.