论文部分内容阅读
【目的】针对电子易货(e-Barter)这一新兴C2C在线交易模式,提出优于其现有(1,0,–1)评分制的多维信誉模型e-BRM。【方法】e-BRM基于Wilson评分区间计算易货者好评率,基于等概率分布计算易货者好评覆盖率,并通过时效衰减因子、差评惩罚因子、实名认证因子等指标实现对易货者交易值的聚合处理。【结果】e-BRM最终将得到的三元组<好评率,覆盖率,交易值>聚合为统一的易货者信誉度,较(1,0,–1)评分制更能表征易货者真实信誉水平。【局限】在实际应用e-BRM时,可单独设计模型的在线增量更新机制以改善实时性。【结论】仿真实验结果能够证明e-BRM模型的有效性,电子易货交易双方可据此做出合理交易决策以降低交易风险。
【Objective】 In view of e-Barter, an emerging C2C online trading model, multi-dimensional credit model e-BRM is proposed which is superior to the existing (1,0, -1) scoring system. 【Method】 e-BRM calculates bumper rates based on the Wilson score interval, and calculates bumper coverage based on equal probability distribution. E-BRM achieves the goal of bartering by means of aging attenuation factor, bad judgment penalty factor and real name authentication factor Aggregation of transaction values. 【Result】 The e-BRM finally aggregates the obtained trio into a unified bachelor’s credit rating, which can better characterize the barter than the (1,0, -1) True credit level. [Limitations] In the practical application of e-BRM, the model can be designed online incremental update mechanism to improve real-time. 【Conclusion】 The simulation results can prove the validity of the e-BRM model. Both sides of the electronic barter can make reasonable trading decisions to reduce the transaction risk.