基于置信区间进行POS机异常检测算法的设计

Design of Anomaly Detection Algorithm of POS by Confidence Interval Prediction

为保证校园卡数据一致性,设计了一种基于置信区间的POS机流水异常检测算法。利用此算法可以得知POS机流水上传功能的运行情况,能及时发现流水未及时上传的POS机并发出预警,从而确保消费者消费金额及时扣除,商户收入及时入账,有效避免账务不平、流水丢失等问题发生,对于校园卡系统的稳定运行,对于维护消费者和商户的利益具有十分重要的意义。随着消费数据的累积,及时对算法进行修正,POS机异常的判断准确率会有较大提升。

In order to ensure the consistency of campus card data,an algorithm of Pos flow detection was designed based on confidence interval.Using the algorithm,we can know the running sitution of flow upload function,find the POS whose was not uploaded in time,so as to ensure the consumer’s consumption amount was deducted,and the merchant’s income was recorded in the account,which can effectively avoid the problems such as uneven accounting and loss of flow.it is very importance for the stable operation of campus card and the protection of the interest of consumers and merchants.Modify the algorithm with the accumulation of consumption data,the accuracy will be greatly improved.