论文部分内容阅读
利用交易评价数据对商品和卖家进行信用评价以供用户参考成为电子商务在线交易平台最基本的服务。然而,目前的信用评价方法很少考虑不同用户间的评价偏好差异,将所有用户的评价同等看待,导致蓄意刷分或恶意差评等信用造假问题屡禁不止。该文提出了一种基于群体偏好的交易评价可信度确立方法。首先采用K-means聚类算法将用户分为3类用户群,通过实证数据分析验证了用户群间明显的评价偏好差异,然后利用评价偏好特征确立每类用户不同类型交易评价的可信度,并提出了动态的交易评价可信度更新策略。该方法能够有效地限制信用造假行为。
Use the transaction evaluation data to evaluate the goods and sellers for the user’s reference to become the most basic service of the e-commerce online trading platform. However, the current credit evaluation methods seldom consider the differences in evaluation preferences among different users, and treat all users as equally evaluated, resulting in repeated fraud and malicious negative ratings. This paper presents a method to establish the credibility of transaction evaluation based on group preference. Firstly, K-means clustering algorithm is used to classify users into three types of user groups. Empirical data analysis verifies the obvious differences in evaluation preferences between user groups. Then, using the evaluation preference characteristics, the credibility of different types of transactions for each type of users is established. And put forward a dynamic strategy to update credibility of transaction evaluation. This method can effectively limit the credit fraud behavior.