A new compression based method for android malware detection using opcodes

Nowadays, the functionality of mobile devices improved substantially which in some cases they were as capable as personal computers. We perform a wide range of our daily tasks with mobile devices like browsing the internet, checking mail, social networking and transforming money. As these smart devices become more popular and usable, they attracted more attackers. Recently, mobile malwares increased sharply and their caused detriments menace the usability and privacy due to the sensitive data which are stored in these devices. According to the intense increase in the number of these attacks yearly, malware detection becomes a prominent topic in mobile security. Since traditional signature based techniques which are used by commercial anti-virus have failed to detect new and obfuscated malwares, machine learning approaches have been employed to find and detect behavior patterns of malwares from extracted features. In this paper, a new heuristic malware detection technique was proposed based on compression methods. The momentous superiority of this approach is using opcode as an input for compression models which causes accuracy to be increased. To assess the potency of the proposed methods, several experiments are conducted. The experimental results of method show promising improvement of accuracy to support the main idea.