Multi resolution analysis of anomaly detection based on intelligent card consumption data

In the process of using smart card, it may cause the smart card machine to upload to the background wrong or missing data, also caused economic losses to the enterprise or business as a result of the fault or fault line card in the machine and other factors. To solve the data transmission process to improve the data upload from the performance angle problem. Firstly using the wavelet multi-resolution analysis to get result of customer consumption data of high frequency and low frequency separation. Secondly, using the least square method (multivariate linear regression) the difference between the original signal and the reconstructed signal processing, find out the abnormal data in the record of customer consumption. Detection of abnormal data customer consumption and demonstrate that the method can better by testing the school supermarket consumption record. For the enterprises to financial workers or merchants to provide more reliable data for consumption.