一种抗欺诈的C2C卖方信誉计算模型研究

Fraud Against Model for Seller's Reputation of C2C

针对C2C信誉模型中小额商品信誉炒作、信誉共谋、信誉诋毁等问题,引入交易价格、反馈可信度、共谋因子等参数,提出一种买方视角下抗欺诈的卖方成员信誉计算模型(C2CRep)。实验中通过收集网络交易数据,定义可疑欺诈的基本特征对数据进行抽取,并设定信誉计算误差(RCE)指标检验由欺诈行为带来的信誉值在社区信誉所占比例来检验模型的应用效果。结果表明,C2CRep在3类不同比例的欺诈行为中,RCE明显低于SPORAS与淘宝信誉模型,且RCE值在3类实验中都低于15%,抗欺诈性强。

A fraud against model for seller＇s reputation（C2CRep） from the buyer＇s view was put forward through introducing the price,credibility of feedback,collusion to resist fraud by abundance of cheap goods,collusion,and slander problems.The online purchase and feedback data was used to verify the model＇s validity.The suspicious fraud data was filtered by basic characteristics defined,and reputation compute error index（RCE） was presented.Results show that RCE of C2CRep is significantly lower than SPORAS and Taobao＇s reputation model in three type fraud.RCE is below 15% under three situations,and it reflects a strong anti-attack performance.