Research On Malicious Behavior Detection Of Mobile Payment In Android System

Android system's market share ranked first in all mobile operating system.However,due to the chaotic distribution of the Android application market,users are not vigilant about their security issues,which makes species and number of Android malicious applications extremely rapidly increasing.One of the most representative services is mobile payment,more and more users choose to pay in the mobile terminal,at the same time,the security of mobile payment is also increasingly prominent.Therefore,it is very important to detect the malicious behavior of mobile payment in Android system.It has always been a hot topic in the field of mobile security.Aiming at the problem of mobile payment security which is urgently needed to be solved,the mobile phone operating system is taken as the research platform,and based on the M-D-H dynamic analysis,a mobile payment-based mobile payment is proposed based on the characteristic data of the related API and API call sequence which is called by the malicious behavior of mobile payment Malicious detection model.The main research work of this paper is as follows:(1)In order to fully run the sample to find potentially more malicious behavior,an improved M-D-H dynamic analysis method based on real device-based hybrid test input is proposed to solve the problem of using real device based on random input generation method Wi-Fi off or the device is set to flight mode,resulting in the sample can not run completely and thus can not find more malicious behavior problems.Through a large number of comparative experiments to verify the M-D-H method,the author find more potential malicious behavior indicated by the API call.(2)Aiming at the ineffectiveness of commonly used feature selection methods for feature extraction of API call sequences,an optimized feature extraction method based on cyclic high-order Markov reduction model is proposed.This method can extract the API sequence with high recognition degree and combine it with the mobile payment malicious API to construct the feature database in this paper.Making the extracted feature data will not produce sparseness and will not appear redundant phenomenon,improve the detection efficiency of the model.(3)Propose mobile payment malicious detection model.The optimized M-D-H dynamic method is used to obtain the characteristic data of the sample,and then the feature extraction is performed by using the optimized high-order Markov reduced-order model.Finally,classification is performed by using the classification algorithm.Python language and other tools is used to achieve,and verify the validity of the test model through experimental data.