摘要
电子商务环境中商品评论对于客户有极强的导引作用,因此商家往往会制造虚假评论,欺诈消费客户。为了监控商家该种行为,将同款商品在不同商家处的评论,以情感计算的方式,转化为特征向量,与第三方专业评测网站和微博评论转化得到的向量比较,识别其中偏离度大的向量,即该商家的商品可能存在欺诈。实验证明,系统能够识别欺诈商家。
Product comments in e-commerce environment has sqrong guidance role on customers, therefore the merchants will usually falsify bogus comments to defraud customers. In order to monitor such behaviours of merchants, we transform the comments of the same product from different merchants to eigenvector by the way of affective computing, then compare it with the vector derived from the comments on the third-party professional assessment and test websites as well as on the microblogging, and identify those vectors with big deviation degree, that means the product of those merchants are possibly in fraudulence. Experiment proves that the system can recognise fraudulent merchants.
出处
《计算机应用与软件》
CSCD
2015年第9期54-58,77,共6页
Computer Applications and Software
关键词
离群点
电子商务
微博
情感计算
语义
Outlier detection E-commerce Microblogging Affective computing Semantics
