Abstract

This article mainly studies ATM transaction feature analysis and anomaly detection. Select the trading success rate, transaction response time and other characteristic parameters, analyze the relationship between transaction response time and transaction success rate. After comparing the fuzzy C-means clustering algorithm and the K-means clustering algorithm, K-means clustering algorithm was used to classify the transaction response time. Thendesign trading anomaly detection program. The use of naive Bayesian classifier for data classification can determine the alarm level. And using Gaussian distribution and Laplace smoothing calibration to increase model accuracy to reduce false alarm. The use of MATLAB programming to get the following result: When the transaction response time is between 0 ~ 85.57, the system predicts to be successful. When the transaction response time between 85.57 ~ 212.31, the system predicts a warning. When the transaction response time between 212.31 ~ 1007.8, the system predicts that the alarm.

Authors and Affiliations