| ABSTRACT:In recent years, mobile payment has attracted more users with its convenience and efficiency. However, the security mechanism and technology of the application layer are still not complete. Moreover, this fascinating way of paying is still affected by the emergence of mobile virus and the loopholes in mobile terminals, open banking system and other aspects of the interface, which makes the mobile payment become one of the current research focus.This paper raised problems caused by the malicious behavior of mobile payment. We analyzed data from abnormal traffic monitoring model based on Back Propagation Neural Network(BP neural network) and used the information entropy as the basic measurement method. We conducted a completed analysis of traffic flow and produced a real-time adjustment of the standard entropy through comparing the process of network flow between abnormal and standard behavior in mobile payment. The work can be demonstrated in the following aspects:1. Through analyzing of mobile payment process, we discussed the main sources of security threats. By selecting the characteristic of parameters, which can fully reflect its flow characteristics, we can achieve its traffic monitoring in order to ensure its security.2. This paper designed a structure of abnormal flow monitoring model by using the most prominent re-learning mechanism in BP neural network which can be sent the initial monitoring results again into the neural internal network, in order to achieve the dynamic analysis and adjustment purpose.3. This paper proposed a real-time adjustment algorithm of standard entropy in order to adapt the dynamical network traffic. Furthermore, the algorithm was designed to only select the significant changes in the characteristic parameter, given the efficiency of the monitoring results.Finally, we completed the analysis and tested of this anomaly monitoring system by using MATLAB.In conclusion, the model worked well as a role in resistance the certain malicious network attacks which can be applicable for abnormal traffic monitoring in secure mobile payment.This work has been supported by the National Natural Science Foundation of China under Grant61371071, Beijing Natural Science Foundation under Grant4132057, the Beijing Science and Technology Program under Grant Z121100007612003, and so on. |
