Detection Of Android Malware Based On Program Slicing

With the rapid development of Android smart phones, more and more people suffer from secret deductions, privacy theft and system destruction in recent years. How to deal with the rapid development of Android malware and detect Android malware accurately with fewer resources is a worthy topic. Existing techniques on Android malware detection mainly focus on characteristic code and malicious behavior. Therefore, in this paper we put forward a kind of malware detection technology based on program slicing.Firstly, we study program slicing and calculate the cohesion and coupling degree relying on object-oriented systems dependence graph. Then we take this technology into Android malware detection to form a new safety evaluation method to detect Android malware, including measuring system dependence graph, analyzing privilege characteristic and the DEX file. Finally, through experiments it proves that this detection method is more excellent with high feasibility and effectiveness.The innovation is in the following :(1) Apply program slicing technique to Android malware detection, through the analysis of the cohesion and coupling degree of system dependence graph, to get Android malware security feature.(2) Propose permissions preprocessing algorithm based on information entropy, in order to remove permissions weakly related with identifying malicious software.(3) Put forward a set of feasible and multi-dimension safety evaluation indicators to check security of Android applications, and demonstrate its effectiveness through experiments.