论文部分内容阅读
【目的】针对电子商务推广中出现的共谋虚增销量的欺诈问题,提出一种基于模板用户信息搜索行为和统计分析的共谋销量识别方法。【方法】为了刻画用户在C2C网站购物时的信息搜索行为,提出一种带关键字的用户信息搜索行为模型以及信息搜索行为的相似度计算方法;依据共谋用户信息搜索行为的相似性,提出一种基于层次聚类的欺诈嫌疑挖掘算法;给出基于统计分析的欺诈识别方法从欺诈嫌疑中识别共谋买家,以实现对卖家销售记录中虚增销量的识别。【结果】在改进的数据集上验证该方法的召回率和准确率分别为88.6%和90.1%。【局限】不能动态调整用于识别欺诈嫌疑行为是否为“刷单”的阈值。【结论】该方法可有效识别基于模板用户信息搜索行为的共谋虚增销量。
【Objective】 In order to solve the problem of fraud in consummate virtual sales growth in e-commerce promotion, a method of collusive sales identification based on template user information search behavior and statistical analysis is proposed. 【Method】 In order to characterize the information searching behavior of users when they shop in C2C website, a kind of user information search behavior model with key words and the similarity calculation method of information search behavior are proposed. Based on the similarity of search behavior of collusion users information, A kind of fraud suspicious mining algorithm based on hierarchical clustering is proposed; the method of fraud identification based on statistical analysis is given to identify conspiracy buyers from the fraud suspicion, so as to realize the identification of the virtual sales volume in the sales records of the sellers. 【Result】 The results show that the recall and accuracy of the proposed method are 88.6% and 90.1% respectively on the improved data set. [Limitations] You can not dynamically adjust the threshold used to identify fraud suspects as a “brush order.” 【Conclusion】 This method can effectively identify the conspiracy to increase sales based on the template user information search behavior.